Meeting Title: AI-Default-ABC Standup Date: 2025-11-05 Meeting participants: Samuel Roberts, Casie Aviles, Gabriel Lam, Rico Rejoso, Mustafa Raja, Uttam Kumaran, Awaish Kumar, Demilade Agboola, Zoran Selinger, Amber Lin, Robert Tseng, Henry Zhao
WEBVTT
1 00:00:27.800 ⇒ 00:00:28.510 Casie Aviles: Damn.
2 00:00:30.580 ⇒ 00:00:31.650 Samuel Roberts: Ayy.
3 00:00:32.570 ⇒ 00:00:33.799 Samuel Roberts: How’s it going?
4 00:00:36.010 ⇒ 00:00:37.370 Casie Aviles: Yeah, doing good.
5 00:00:37.870 ⇒ 00:00:38.870 Casie Aviles: How about you guys?
6 00:00:39.340 ⇒ 00:00:40.190 Casie Aviles: Hey, Gabe.
7 00:00:40.810 ⇒ 00:00:42.170 Gabriel Lam: Morning, morning.
8 00:00:42.430 ⇒ 00:00:43.400 Samuel Roberts: Good morning.
9 00:00:46.670 ⇒ 00:00:47.690 Samuel Roberts: How’s everyone?
10 00:00:50.300 ⇒ 00:00:54.179 Gabriel Lam: Pretty good. It’s very cold in the East Coast, but I’m doing good.
11 00:00:54.180 ⇒ 00:00:59.859 Samuel Roberts: I was gonna ask where you still… where you were at this point, because, yeah, I… I feel that this morning here, too, is a little chilly.
12 00:01:03.450 ⇒ 00:01:07.999 Samuel Roberts: Cool. How was everyone’s yesterday?
13 00:01:13.330 ⇒ 00:01:19.180 Mustafa Raja: Good, I mostly worked on the internal stuff, and then, some default.
14 00:01:19.690 ⇒ 00:01:20.290 Samuel Roberts: Cool.
15 00:01:20.830 ⇒ 00:01:22.099 Mustafa Raja: Yeah. How was yours?
16 00:01:23.220 ⇒ 00:01:24.399 Samuel Roberts: Pretty good.
17 00:01:25.280 ⇒ 00:01:28.559 Samuel Roberts: Can you play too much. Yeah, I was following along with some of that stuff, trying to…
18 00:01:28.970 ⇒ 00:01:32.920 Samuel Roberts: things out a little bit, but let’s actually jump right into that, if we can.
19 00:01:32.920 ⇒ 00:01:33.620 Mustafa Raja: Yeah.
20 00:01:34.840 ⇒ 00:01:38.279 Samuel Roberts: So I was just testing, I saw the branch,
21 00:01:39.410 ⇒ 00:01:42.140 Samuel Roberts: Let me… hold on, where is it?
22 00:01:43.410 ⇒ 00:01:46.309 Samuel Roberts: Yeah, I was wondering… so it looks like it is pulling in…
23 00:01:47.240 ⇒ 00:01:51.020 Samuel Roberts: Zoom? It says 3 Zoom stand-ups when I run ABC.
24 00:01:51.680 ⇒ 00:02:04.299 Samuel Roberts: We probably just want to change that to Zoom meetings, because they’re not all stand-ups, I suppose, but that’s a nitpicky thing. But I’m not seeing, like… we don’t have a way that it’s… I think the next thing is figure out, like, how do we cite that, right?
25 00:02:05.020 ⇒ 00:02:05.350 Gabriel Lam: Yeah.
26 00:02:05.350 ⇒ 00:02:05.790 Casie Aviles: Yeah.
27 00:02:05.790 ⇒ 00:02:06.360 Mustafa Raja: Yes.
28 00:02:06.360 ⇒ 00:02:07.830 Samuel Roberts: Because let me, let me share my screen.
29 00:02:07.970 ⇒ 00:02:09.470 Samuel Roberts: Real quick.
30 00:02:12.600 ⇒ 00:02:29.180 Samuel Roberts: Okay, so I just opened this up, just ran it for… re-ran it for ABC two minutes ago. So yeah, I think we can change this to Zoom meetings, but that’s, like I said, not a big deal. I don’t have any weekly goals in here, but it does… it still feels very linear-heavy. At least we got one Slack follow-up, which is good.
31 00:02:29.620 ⇒ 00:02:36.569 Samuel Roberts: But I don’t know, these all seem… Like, they have Slack
32 00:02:36.890 ⇒ 00:02:40.240 Samuel Roberts: This one’s… yeah, everything here seems linear, I mean.
33 00:02:40.400 ⇒ 00:02:41.680 Samuel Roberts: So I’m wondering…
34 00:02:42.920 ⇒ 00:02:44.010 Mustafa Raja: Yeah, what I…
35 00:02:44.010 ⇒ 00:02:44.789 Samuel Roberts: Yeah, go ahead.
36 00:02:44.980 ⇒ 00:02:47.390 Mustafa Raja: What I’ve found is,
37 00:02:48.350 ⇒ 00:03:03.630 Mustafa Raja: In Slack, a lot of the times, we are mentioning tickets, as we share progress on them. Sure, good point. So sometimes it… what it does is it infers, okay, this is coming from Linea and Slack.
38 00:03:03.910 ⇒ 00:03:13.309 Mustafa Raja: And since it does have the link for what’s it called, linear, it displays it in a way that it looks like it’s coming from linear.
39 00:03:13.310 ⇒ 00:03:16.129 Samuel Roberts: Oh, okay, that’s fair, that’s fair.
40 00:03:16.390 ⇒ 00:03:21.800 Mustafa Raja: Yeah, I guess, we still have the issue with Zoom, I don’t see anything.
41 00:03:22.020 ⇒ 00:03:23.070 Samuel Roberts: Yeah, I wonder…
42 00:03:23.070 ⇒ 00:03:24.720 Mustafa Raja: reference to Zoom.
43 00:03:25.940 ⇒ 00:03:30.959 Casie Aviles: It did show up for… I generated… I…
44 00:03:31.160 ⇒ 00:03:35.130 Casie Aviles: Generated the… what do you call this? The status updates for…
45 00:03:36.100 ⇒ 00:03:40.759 Casie Aviles: One of the clients. There is Zoom, but the point issue there is…
46 00:03:41.740 ⇒ 00:03:44.439 Casie Aviles: For the AI team, can you try for the AI team?
47 00:03:44.740 ⇒ 00:03:45.240 Samuel Roberts: Sure.
48 00:03:45.240 ⇒ 00:03:53.229 Casie Aviles: Like, one of the issues there that I found was it got, like, an item from a different client.
49 00:03:53.910 ⇒ 00:03:58.420 Casie Aviles: Look at, yeah, the one for Zoran. Yeah, for Zoran at the bottom.
50 00:03:58.420 ⇒ 00:03:59.180 Samuel Roberts: Yeah…
51 00:03:59.180 ⇒ 00:03:59.720 Casie Aviles: 12.
52 00:04:00.010 ⇒ 00:04:04.049 Casie Aviles: So this should not be here, I think that’s something I should work on.
53 00:04:04.820 ⇒ 00:04:10.840 Samuel Roberts: Yeah, definitely. That’s gonna be a… an ongoing thing we need to sort out anyways, how to…
54 00:04:11.280 ⇒ 00:04:12.570 Samuel Roberts: Split those up a bit.
55 00:04:12.720 ⇒ 00:04:14.370 Samuel Roberts: Okay.
56 00:04:14.750 ⇒ 00:04:18.790 Samuel Roberts: What is… is this… is this a real ticket that it’s linking to here?
57 00:04:19.740 ⇒ 00:04:26.569 Mustafa Raja: Yeah, okay, so that definitely seems, that seems very high, yeah.
58 00:04:28.130 ⇒ 00:04:29.159 Rico Rejoso: We’re eating.
59 00:04:31.020 ⇒ 00:04:32.310 Mustafa Raja: Yeah, but didiano.
60 00:04:32.600 ⇒ 00:04:33.100 Mustafa Raja: AI.
61 00:04:34.660 ⇒ 00:04:36.200 Mustafa Raja: Oh, weird.
62 00:04:36.200 ⇒ 00:04:38.089 Samuel Roberts: So I definitely did something weird, kinda.
63 00:04:38.090 ⇒ 00:04:39.439 Mustafa Raja: hallucination there, maybe?
64 00:04:40.070 ⇒ 00:04:40.730 Mustafa Raja: Yeah.
65 00:04:41.330 ⇒ 00:04:42.010 Samuel Roberts: Okay.
66 00:04:42.510 ⇒ 00:04:45.940 Casie Aviles: And, yeah, it assigned a wrong ticket, I guess.
67 00:04:45.940 ⇒ 00:04:46.970 Samuel Roberts: Yeah.
68 00:04:47.490 ⇒ 00:04:48.400 Samuel Roberts: Okay.
69 00:04:48.810 ⇒ 00:04:54.250 Samuel Roberts: Let’s, I’m wondering what the best way to tackle this is, then.
70 00:04:58.940 ⇒ 00:05:04.160 Samuel Roberts: Should we just step through and see how it’s, like, talk through it a little bit, maybe? Would that be helpful?
71 00:05:06.900 ⇒ 00:05:08.229 Casie Aviles: Or, or…
72 00:05:08.230 ⇒ 00:05:09.150 Samuel Roberts: Yeah.
73 00:05:10.970 ⇒ 00:05:11.979 Samuel Roberts: Sorry, go ahead.
74 00:05:13.030 ⇒ 00:05:16.519 Casie Aviles: Yeah, I was just thinking about what’s the best way to…
75 00:05:16.780 ⇒ 00:05:24.240 Casie Aviles: To separate those, because it’s really… it’s a bit tricky, since everything is in one meeting.
76 00:05:24.450 ⇒ 00:05:28.379 Samuel Roberts: Yeah, that might be something we need to tackle separately, perhaps, then.
77 00:05:28.760 ⇒ 00:05:33.430 Samuel Roberts: Cause we’re gonna need to worry about that.
78 00:05:33.430 ⇒ 00:05:34.060 Mustafa Raja: if…
79 00:05:34.270 ⇒ 00:05:34.890 Samuel Roberts: Go ahead.
80 00:05:35.280 ⇒ 00:05:42.360 Mustafa Raja: I’m wondering if we can have an AI step that, strips the content, particularly to the client.
81 00:05:42.480 ⇒ 00:05:44.730 Mustafa Raja: In terms of the transcript.
82 00:05:44.730 ⇒ 00:05:45.680 Casie Aviles: Yeah.
83 00:05:46.140 ⇒ 00:05:46.960 Samuel Roberts: Right.
84 00:05:47.540 ⇒ 00:05:48.030 Mustafa Raja: But that.
85 00:05:48.030 ⇒ 00:05:49.149 Casie Aviles: Maybe you could do.
86 00:05:52.400 ⇒ 00:05:55.839 Mustafa Raja: That has to be very, very accurate, though.
87 00:05:56.230 ⇒ 00:05:58.349 Samuel Roberts: Yes, it does, cause this is gonna be…
88 00:05:59.510 ⇒ 00:06:01.589 Mustafa Raja: If you’re going to be relying on that task.
89 00:06:01.590 ⇒ 00:06:11.399 Samuel Roberts: Yeah, okay, so let’s… let’s think about that. So that we’re going to need here, but we’re eventually going to need that probably a few other places, because I imagine we’re going to want to pull that in another way.
90 00:06:11.650 ⇒ 00:06:15.999 Samuel Roberts: So I think for now, we can think of it as part of this step, maybe?
91 00:06:16.560 ⇒ 00:06:20.369 Samuel Roberts: But let’s keep it as a separate, like, master agent.
92 00:06:20.850 ⇒ 00:06:24.540 Samuel Roberts: So we could potentially use that somewhere else, but…
93 00:06:24.830 ⇒ 00:06:25.500 Mustafa Raja: Yeah.
94 00:06:26.170 ⇒ 00:06:29.770 Samuel Roberts: So maybe we need to just work on that as an agent, because that would help with some of this stuff.
95 00:06:29.890 ⇒ 00:06:30.830 Samuel Roberts: At least.
96 00:06:31.900 ⇒ 00:06:33.500 Casie Aviles: For classification, right?
97 00:06:33.500 ⇒ 00:06:37.560 Samuel Roberts: Yeah, so we need to figure out… roughly…
98 00:06:37.700 ⇒ 00:06:43.009 Samuel Roberts: well, there’s a couple ways we could do it. We could just try to identify the part, the, like, actual…
99 00:06:43.490 ⇒ 00:06:45.919 Samuel Roberts: Text from the transcript. That is…
100 00:06:46.280 ⇒ 00:06:49.160 Samuel Roberts: Relevant to the client or team, or whatever.
101 00:06:50.330 ⇒ 00:06:56.070 Samuel Roberts: Right? The other way we could do it is by kind of doing that, but building in, like, a…
102 00:06:56.340 ⇒ 00:06:57.450 Samuel Roberts: Timestamp.
103 00:06:57.960 ⇒ 00:06:59.530 Samuel Roberts: Chapters, almost.
104 00:07:01.530 ⇒ 00:07:11.120 Samuel Roberts: The issue with that is we’re not necessarily… we’re kind of consistent in the way we break it up, but we definitely could… jumping around might still happen in a meeting.
105 00:07:11.610 ⇒ 00:07:12.470 Mustafa Raja: Yeah.
106 00:07:14.130 ⇒ 00:07:18.659 Samuel Roberts: So, I think we probably want something that’s just gonna pull all the relevant stuff from the meeting.
107 00:07:19.020 ⇒ 00:07:19.920 Samuel Roberts: Right
108 00:07:24.100 ⇒ 00:07:28.249 Samuel Roberts: Because the chapter… like, chapters would be kind of cool, because if you came to, like…
109 00:07:29.080 ⇒ 00:07:37.869 Samuel Roberts: a meeting here. You know, if we went to, like, what do we got?
110 00:07:38.960 ⇒ 00:07:40.320 Samuel Roberts: Awesome.
111 00:07:41.140 ⇒ 00:07:50.059 Samuel Roberts: like, if we came here… how long is this? This one is 54 minutes. We could do something where it’s, like, chapters, and you could jump around by client or something, but we don’t.
112 00:07:50.060 ⇒ 00:07:53.070 Mustafa Raja: One more thing is, yesterday,
113 00:07:53.450 ⇒ 00:08:02.430 Mustafa Raja: two meetings, two stand-ups that we have, have the same name. I think it is because we’re using the same Zoom link.
114 00:08:02.950 ⇒ 00:08:04.919 Mustafa Raja: For both of the stand-ups.
115 00:08:05.920 ⇒ 00:08:06.739 Mustafa Raja: If you see…
116 00:08:06.740 ⇒ 00:08:07.720 Samuel Roberts: These guys?
117 00:08:07.990 ⇒ 00:08:08.760 Mustafa Raja: Yeah.
118 00:08:10.620 ⇒ 00:08:12.480 Samuel Roberts: Yeah, 18 minutes… okay.
119 00:08:12.880 ⇒ 00:08:17.269 Samuel Roberts: Okay, yeah, so that’s another thing we gotta start, because this title is being generated.
120 00:08:18.690 ⇒ 00:08:23.509 Mustafa Raja: This is… I think this is coming from Zoom, it’s… Sorry, the calendar itself.
121 00:08:23.850 ⇒ 00:08:26.040 Samuel Roberts: Okay, so because we’re using the same link.
122 00:08:26.300 ⇒ 00:08:27.280 Mustafa Raja: Yeah, yeah.
123 00:08:27.280 ⇒ 00:08:30.059 Samuel Roberts: Yeah. So it definitely…
124 00:08:31.220 ⇒ 00:08:34.399 Mustafa Raja: Let me double-check, though, if it’s exactly the same one.
125 00:08:35.049 ⇒ 00:08:40.079 Samuel Roberts: It did do… I can’t…
126 00:08:41.559 ⇒ 00:08:43.899 Samuel Roberts: These are different IDs, at least, right?
127 00:08:45.120 ⇒ 00:08:47.340 Samuel Roberts: Yeah, so that’s something. Okay.
128 00:08:49.390 ⇒ 00:08:56.749 Samuel Roberts: I don’t really want to tackle that right now, because that’s a little… that’s a whole other thing, but, like, let’s think about that at some point, because that is…
129 00:08:57.920 ⇒ 00:09:02.320 Samuel Roberts: gonna be a problem. But for now, I think we just wanna say, like, let’s look at this meeting.
130 00:09:03.080 ⇒ 00:09:09.709 Samuel Roberts: figure out… The relevant content to whatever client it is.
131 00:09:10.170 ⇒ 00:09:11.040 Samuel Roberts: Correct?
132 00:09:12.540 ⇒ 00:09:18.590 Samuel Roberts: So… That’s like a… let’s call that a pre-step for those, like.
133 00:09:19.150 ⇒ 00:09:26.579 Samuel Roberts: Actually, do I have… I don’t know if I have… oh, I’m not even sharing it, never mind. I was gonna try to walk through the code, but…
134 00:09:29.400 ⇒ 00:09:34.299 Samuel Roberts: let’s just talk about it in the abstract, I guess. So right now, we pull… how do we… how are we pulling the linear…
135 00:09:34.780 ⇒ 00:09:40.119 Samuel Roberts: transcripts? Is it just by… Title, or is it by classification, or…
136 00:09:41.340 ⇒ 00:09:42.510 Casie Aviles: Linear or Zoom?
137 00:09:42.900 ⇒ 00:09:45.030 Samuel Roberts: I’m sorry, Zoom, Zoom, my mistake, thank you.
138 00:09:47.470 ⇒ 00:09:49.039 Casie Aviles: Yeah, for Zoom,
139 00:09:49.410 ⇒ 00:09:57.629 Casie Aviles: We’re just getting it from Superbase, we’re just getting the transcript, and we’re filtering it by the name, or the topic, the meeting name.
140 00:09:57.930 ⇒ 00:10:01.200 Casie Aviles: Okay. So as long as it has, like, stand-up.
141 00:10:01.820 ⇒ 00:10:09.289 Casie Aviles: In the title, and different variations of the word standout, then it’s gonna be pulled, and…
142 00:10:09.420 ⇒ 00:10:14.960 Casie Aviles: It has to look at the meeting date as well, and then it only gets, like, from the previous day.
143 00:10:15.200 ⇒ 00:10:18.120 Casie Aviles: So I guess if a meeting is done.
144 00:10:18.630 ⇒ 00:10:20.880 Casie Aviles: For today, it’s still not gonna show up.
145 00:10:21.160 ⇒ 00:10:21.550 Mustafa Raja: Yeah.
146 00:10:21.550 ⇒ 00:10:23.649 Casie Aviles: Unless it was yesterday’s.
147 00:10:23.980 ⇒ 00:10:24.560 Casie Aviles: Meeting.
148 00:10:24.560 ⇒ 00:10:25.210 Samuel Roberts: Okay.
149 00:10:25.340 ⇒ 00:10:30.700 Samuel Roberts: So, couple things there. I think we probably want it to be more than just stand-ups, right?
150 00:10:32.510 ⇒ 00:10:39.890 Samuel Roberts: Because we’re going to want anything related to whatever client it is. Let me go… Like, if there’s…
151 00:10:40.650 ⇒ 00:10:48.220 Samuel Roberts: you know, AI… if we have a meeting that’s classified as an AI thing, if we got on to, like, a working session, that would be useful information here, too, right?
152 00:10:49.870 ⇒ 00:10:50.730 Casie Aviles: Yeah.
153 00:10:50.880 ⇒ 00:10:51.570 Samuel Roberts: shall we?
154 00:10:51.820 ⇒ 00:10:53.890 Samuel Roberts: Maybe need to think that through a little bit.
155 00:10:54.240 ⇒ 00:10:57.880 Samuel Roberts: I think we also… want…
156 00:10:59.040 ⇒ 00:11:01.179 Samuel Roberts: Oh, and I had another thought beyond that, but…
157 00:11:04.210 ⇒ 00:11:07.470 Samuel Roberts: And then we wanna… yeah, okay, so… So there’s…
158 00:11:08.120 ⇒ 00:11:11.949 Samuel Roberts: I don’t want to really think of them as two different kinds of meetings, but the stand-up meeting
159 00:11:12.200 ⇒ 00:11:16.250 Samuel Roberts: Is the one that we have to break apart into multiple pieces, right?
160 00:11:16.700 ⇒ 00:11:17.450 Casie Aviles: Yes.
161 00:11:18.180 ⇒ 00:11:26.630 Samuel Roberts: Okay, well… One thing I’m noticing here is that, like, even this one, it’s put in department, engineering.
162 00:11:26.730 ⇒ 00:11:29.550 Samuel Roberts: Doesn’t have a team associated, right?
163 00:11:31.790 ⇒ 00:11:38.979 Samuel Roberts: Because before, you could only have one, I think, right? Or now you can only have one. And that was kind of in line with…
164 00:11:40.120 ⇒ 00:11:42.910 Samuel Roberts: How we were running the calendar events.
165 00:11:43.930 ⇒ 00:11:44.650 Mustafa Raja: Yeah.
166 00:11:45.070 ⇒ 00:11:46.929 Samuel Roberts: So do we need to now make this…
167 00:11:47.300 ⇒ 00:11:52.460 Samuel Roberts: that actually make this work? Does this need to be multi… Multiselect?
168 00:11:54.370 ⇒ 00:11:55.239 Mustafa Raja: I guess we…
169 00:11:55.240 ⇒ 00:11:55.870 Casie Aviles: Yeah…
170 00:11:56.800 ⇒ 00:12:00.699 Samuel Roberts: Because I’m… what I’m thinking is, and this is… I’m kind of getting a little ahead of myself, but…
171 00:12:01.080 ⇒ 00:12:06.240 Samuel Roberts: pulling stand-ups, We want to pull, actually, anything related to AI, right?
172 00:12:06.640 ⇒ 00:12:08.610 Mustafa Raja: Or anything related to ABC.
173 00:12:09.880 ⇒ 00:12:17.660 Samuel Roberts: Yeah. So, we want… Like, in an ideal world, this would pull… You know, anything that’s got…
174 00:12:18.250 ⇒ 00:12:21.769 Samuel Roberts: either ABC in the title, or classified as ABC,
175 00:12:23.550 ⇒ 00:12:28.600 Samuel Roberts: I’m just trying to think if there’s a better way to… do that.
176 00:12:28.730 ⇒ 00:12:34.850 Samuel Roberts: without… having to reinvent everything about how we classify meetings. But maybe we just need to do that anyway.
177 00:12:35.890 ⇒ 00:12:39.219 Samuel Roberts: But I don’t want that to be part of this project, if we can help it, you know?
178 00:12:39.220 ⇒ 00:12:39.940 Gabriel Lam: Yoon.
179 00:12:43.200 ⇒ 00:12:49.149 Gabriel Lam: Do you guys think it’s necessary for this sprint? It sounds like there’s gonna be a lot of…
180 00:12:49.390 ⇒ 00:12:49.970 Casie Aviles: Yeah.
181 00:12:49.970 ⇒ 00:12:50.970 Gabriel Lam: Rebuilding.
182 00:12:51.430 ⇒ 00:13:01.989 Samuel Roberts: Yeah, well, there’s already, some rebuilding that needs to happen to get some of this stuff into Mastra, which I’ve been thinking about a bit, and we want to put… I need to put, like, more of a concrete plan together, because I was thinking about the actual…
183 00:13:02.360 ⇒ 00:13:06.080 Samuel Roberts: client hub stuff moving into…
184 00:13:07.340 ⇒ 00:13:11.229 Samuel Roberts: The platform more, but the ingestion and the classification
185 00:13:11.530 ⇒ 00:13:13.650 Samuel Roberts: I haven’t necessarily put thought into.
186 00:13:14.120 ⇒ 00:13:18.099 Samuel Roberts: Moving from N8N into code.
187 00:13:21.000 ⇒ 00:13:29.430 Mustafa Raja: Also for AI channel, we are not… I don’t think we are storing the… what’s it called? The Slack messages in database at all.
188 00:13:32.240 ⇒ 00:13:37.889 Mustafa Raja: So it would say one Slack message, but it, what it is, is, it couldn’t find anything.
189 00:13:37.890 ⇒ 00:13:41.059 Samuel Roberts: Yep, exactly, yeah, I think you’re right. So that’s def… okay, so let’s…
190 00:13:41.910 ⇒ 00:13:45.840 Samuel Roberts: Let’s make a list, I guess, let’s start there. Okay, so what is…
191 00:13:46.910 ⇒ 00:13:48.240 Samuel Roberts: Let’s back up a little bit.
192 00:13:48.460 ⇒ 00:13:50.059 Samuel Roberts: Forget about all the, like.
193 00:13:50.390 ⇒ 00:13:54.250 Samuel Roberts: nitty-gritty that we just talked about. What is not working about some of this right now?
194 00:13:54.540 ⇒ 00:14:00.520 Samuel Roberts: Right? So, Slack messages for internal, probably. I imagine that’s probably consistent on a few.
195 00:14:00.560 ⇒ 00:14:02.309 Mustafa Raja: So we need to start moving…
196 00:14:02.830 ⇒ 00:14:05.009 Samuel Roberts: Slack messages for internal.
197 00:14:06.990 ⇒ 00:14:12.340 Samuel Roberts: We need to start… let me,
198 00:14:13.220 ⇒ 00:14:15.840 Samuel Roberts: Can I ask someone to keep a list, running list as I say, as I go through this?
199 00:14:15.840 ⇒ 00:14:16.389 Gabriel Lam: I got it.
200 00:14:16.780 ⇒ 00:14:26.530 Samuel Roberts: Perfect, thank you. Okay, I didn’t want to just assume, but I want to… I didn’t wanna… Okay. We need to… hold on, this is way again, why do I keep doing this? We need to…
201 00:14:26.720 ⇒ 00:14:29.839 Samuel Roberts: figure out the Zoom issue, right?
202 00:14:30.150 ⇒ 00:14:33.180 Mustafa Raja: Which is a couple things, which I think is…
203 00:14:34.360 ⇒ 00:14:36.739 Samuel Roberts: All client-related meetings.
204 00:14:38.550 ⇒ 00:14:43.029 Samuel Roberts: And then also parsing the stand-up meetings.
205 00:14:43.170 ⇒ 00:14:44.859 Samuel Roberts: Does that sound? I’m wondering…
206 00:14:45.180 ⇒ 00:14:48.640 Mustafa Raja: I’m wondering if we can use turbo buffer for this stuff.
207 00:14:49.890 ⇒ 00:14:55.530 Samuel Roberts: Good suggestion. Yes, Turbo Puffer, we’re already doing something with the search, and it seems pretty quick.
208 00:14:56.120 ⇒ 00:14:56.960 Mustafa Raja: Yeah.
209 00:14:57.760 ⇒ 00:14:58.330 Samuel Roberts: Right.
210 00:14:58.810 ⇒ 00:14:59.969 Samuel Roberts: That’s a good call.
211 00:14:59.970 ⇒ 00:15:05.299 Mustafa Raja: But we’ll, I guess we’ll have to, look if it’s… if it’s feasible for this or not.
212 00:15:05.740 ⇒ 00:15:13.710 Samuel Roberts: Right, okay, so what would that mean? It means we would want to do a quick search over the last 24 hours for anything related to the client.
213 00:15:13.710 ⇒ 00:15:16.010 Mustafa Raja: Okay, yeah, yeah, this is what I’m thinking.
214 00:15:16.140 ⇒ 00:15:21.620 Samuel Roberts: Okay, that’s… that’s a good idea. That might solve some of the problems for Zoom, then. So we would basically…
215 00:15:22.110 ⇒ 00:15:26.849 Samuel Roberts: Correct me if I’m wrong here, we do that search in Turbo Puffer, That would return…
216 00:15:27.110 ⇒ 00:15:33.200 Samuel Roberts: the list of meetings and the SuperBase, that’s what I mean, yeah, the list of meetings we’d pull from Superbase to get the transcripts for.
217 00:15:33.510 ⇒ 00:15:38.620 Mustafa Raja: Yeah, we get the transcripts, and then we, partition only the parts.
218 00:15:38.920 ⇒ 00:15:40.899 Mustafa Raja: Where we talk about the client.
219 00:15:41.520 ⇒ 00:15:48.630 Mustafa Raja: Okay. And all of this needs to happen in 30 seconds, because, if it… if it goes above 30 seconds.
220 00:15:48.860 ⇒ 00:15:53.230 Mustafa Raja: Hiroku… Hiroku doesn’t let us, do that.
221 00:15:53.620 ⇒ 00:16:02.939 Samuel Roberts: Yes, I had… I was wondering about that. I saw that in one of the commits. I think what we need to do here… that may be another… this is another thing now, is…
222 00:16:03.130 ⇒ 00:16:08.590 Samuel Roberts: perhaps when I click this refresh, we can… Make it multi-step here.
223 00:16:09.250 ⇒ 00:16:13.440 Samuel Roberts: Because I think if you respond… Wickly?
224 00:16:14.640 ⇒ 00:16:16.120 Samuel Roberts: It keeps it open longer.
225 00:16:16.570 ⇒ 00:16:17.220 Mustafa Raja: Yep.
226 00:16:17.610 ⇒ 00:16:19.500 Samuel Roberts: So, if we can send back…
227 00:16:19.880 ⇒ 00:16:24.110 Samuel Roberts: And maybe this means we have to break the… the Mastra steps into multiple ones?
228 00:16:25.320 ⇒ 00:16:26.090 Mustafa Raja: That’s weird.
229 00:16:26.090 ⇒ 00:16:34.650 Samuel Roberts: But I think if we could, like, click refresh data, and it would be, like, fetching Zoom, fetching Slack, fetching linear, and it’s sending anything back, I think it will hold that connection open.
230 00:16:35.750 ⇒ 00:16:38.230 Samuel Roberts: Is it Heroku that times out the connection? It’s not the actual…
231 00:16:38.230 ⇒ 00:16:39.030 Mustafa Raja: It is moving forward.
232 00:16:39.030 ⇒ 00:16:41.979 Samuel Roberts: function, right? It’s just… it’s the connection that’s timing out, not the, like.
233 00:16:42.060 ⇒ 00:16:42.790 Mustafa Raja: Mmm.
234 00:16:42.790 ⇒ 00:16:44.969 Samuel Roberts: It’s not like a serverless kind of function or anything.
235 00:16:44.970 ⇒ 00:16:46.810 Mustafa Raja: API the times out.
236 00:16:47.760 ⇒ 00:16:50.410 Samuel Roberts: the API for…
237 00:16:51.320 ⇒ 00:17:00.600 Mustafa Raja: So when we… when we hit the… when we hit the refresh data button, we hit an API, and that particular API cannot run for more than 30 seconds.
238 00:17:01.050 ⇒ 00:17:06.059 Samuel Roberts: I think it… I think it can’t not respond for more than 30 seconds.
239 00:17:06.069 ⇒ 00:17:06.939 Mustafa Raja: Yeah, yeah.
240 00:17:06.940 ⇒ 00:17:08.969 Samuel Roberts: Yeah, that’s what I’m saying, it can run longer.
241 00:17:10.050 ⇒ 00:17:10.410 Mustafa Raja: need to respond.
242 00:17:10.410 ⇒ 00:17:11.700 Samuel Roberts: respond sooner.
243 00:17:12.560 ⇒ 00:17:25.810 Samuel Roberts: With, like, a received kind of thing. I think I understand that. Maybe I’ll take that today and look at that a little bit more, because I was reading about this before we had some Heroku stuff. Okay, so we’ve got, Slack messages need to get in for internal.
244 00:17:27.640 ⇒ 00:17:31.470 Samuel Roberts: Zoom… Perhaps using Turbo Puffer.
245 00:17:31.650 ⇒ 00:17:33.890 Samuel Roberts: And then the Heroku timeout thing.
246 00:17:34.650 ⇒ 00:17:42.430 Samuel Roberts: Right? Okay, so, are there any other outstanding things we can think about here? Obviously, like, that’s gonna affect
247 00:17:42.640 ⇒ 00:17:46.510 Samuel Roberts: What’s showing up here, hopefully, so that hopefully solves a couple of these problems.
248 00:17:46.780 ⇒ 00:17:47.740 Samuel Roberts: But…
249 00:17:47.890 ⇒ 00:17:56.459 Samuel Roberts: We need to figure that out with the Zoom, especially splitting up, and the Slack messages getting in. That should hopefully help keep this a little bit more,
250 00:17:58.110 ⇒ 00:18:01.339 Samuel Roberts: useful overall, and not just linear-focused. Hey, Tom.
251 00:18:01.340 ⇒ 00:18:02.739 Uttam Kumaran: Hey guys, good morning.
252 00:18:02.740 ⇒ 00:18:03.340 Mustafa Raja: Morning.
253 00:18:03.340 ⇒ 00:18:04.360 Samuel Roberts: Good morning.
254 00:18:05.740 ⇒ 00:18:06.640 Uttam Kumaran: That’s weird.
255 00:18:07.240 ⇒ 00:18:07.990 Uttam Kumaran: Yeah, go ahead.
256 00:18:08.380 ⇒ 00:18:16.819 Samuel Roberts: No, no, you’re good, you’re good, you’re good. We were just walking through some stuff, trying to figure out some of the outstanding issues we have, talking it through.
257 00:18:16.980 ⇒ 00:18:22.069 Uttam Kumaran: Cool, looks good. Design looks, looks, better, more dense.
258 00:18:22.290 ⇒ 00:18:22.950 Samuel Roberts: Yeah.
259 00:18:23.080 ⇒ 00:18:30.109 Samuel Roberts: Yeah, so this was the front end that I think Mustafa had worked on while Casey was working on the back end. Casey had put together the one you tested just to…
260 00:18:30.310 ⇒ 00:18:43.829 Samuel Roberts: get some output, so you could see it. This was what we then stitched together. Mustaco stitched those two together. But you can see here, like, it’s only pulling one Slack message from AI, and I think that’s… we think that’s because it’s not actually… we’re not storing the Slack messages.
261 00:18:44.190 ⇒ 00:18:46.360 Samuel Roberts: For the internal channels?
262 00:18:47.140 ⇒ 00:18:54.540 Samuel Roberts: So we gotta add that, because we want those. And then you can see here, I pulled 3 Zoom stand-ups, and that’s a… another issue we’re having, because…
263 00:18:54.760 ⇒ 00:18:57.610 Samuel Roberts: Like, this is… there are two,
264 00:18:58.000 ⇒ 00:19:02.060 Samuel Roberts: With the same title, and this is related to using the same links.
265 00:19:03.260 ⇒ 00:19:04.220 Samuel Roberts: I think.
266 00:19:04.610 ⇒ 00:19:19.320 Samuel Roberts: But the solution we have right now, because I think at some point we need to maybe figure that out, but I don’t want to make that… I don’t want to fold in a bunch of other issues into this. We still have the idea to use Turbo Puffer to do this search, because the other side of this Zoom issue is that
267 00:19:19.630 ⇒ 00:19:22.169 Samuel Roberts: Which one was it that had the wrong…
268 00:19:23.010 ⇒ 00:19:23.920 Casie Aviles: AI team.
269 00:19:23.920 ⇒ 00:19:26.269 Samuel Roberts: Was it AI? Okay, let me go back to that.
270 00:19:26.570 ⇒ 00:19:27.389 Samuel Roberts: Alright, well, I already…
271 00:19:27.390 ⇒ 00:19:28.750 Mustafa Raja: Sunrise around.
272 00:19:28.750 ⇒ 00:19:34.169 Samuel Roberts: Yeah, look at this, the follow-ups here. This is not right. And I think it’s because it’s pulling… it’s not…
273 00:19:34.930 ⇒ 00:19:41.439 Samuel Roberts: getting… it’s… we’re feeding the whole media into context, I think, and rather… and hoping that it parses it properly.
274 00:19:41.890 ⇒ 00:19:46.519 Samuel Roberts: So I think what we need to do is pre-parse it and pass it in.
275 00:19:47.120 ⇒ 00:19:49.120 Samuel Roberts: Are those all the things I talked about?
276 00:19:51.120 ⇒ 00:19:52.420 Samuel Roberts: Am I missing anything on that list?
277 00:19:54.020 ⇒ 00:19:56.729 Gabriel Lam: Just the Heroku thing to look into.
278 00:19:56.730 ⇒ 00:20:07.939 Samuel Roberts: Yes, and the other thing is that we were talking about, there’s a timeout issue with Heroku, but, and if we go to Turbo Puffer and have to then fetch, it might timeout, but I think there’s a solution for that. So hopefully, you know.
279 00:20:08.540 ⇒ 00:20:14.310 Samuel Roberts: Sort that out today, and then we can get to another round of getting this, you know, this more accurate.
280 00:20:14.620 ⇒ 00:20:20.010 Samuel Roberts: More than just linear, but yeah, that’s kind of where we’re at.
281 00:20:21.060 ⇒ 00:20:25.489 Uttam Kumaran: And then how is, tell me about the persistence issues.
282 00:20:25.490 ⇒ 00:20:26.470 Samuel Roberts: Yes.
283 00:20:26.620 ⇒ 00:20:32.730 Samuel Roberts: I believe, this all persists now. Yep, so, and even weekly goals.
284 00:20:32.990 ⇒ 00:20:33.600 Uttam Kumaran: Okay.
285 00:20:33.890 ⇒ 00:20:38.480 Samuel Roberts: Should persist on refresh. Where is that… is that just persisting in the browser still?
286 00:20:38.480 ⇒ 00:20:39.240 Mustafa Raja: Yes.
287 00:20:39.400 ⇒ 00:20:40.200 Uttam Kumaran: Okay.
288 00:20:41.410 ⇒ 00:20:42.770 Samuel Roberts: Go ahead.
289 00:20:42.770 ⇒ 00:20:45.760 Mustafa Raja: And we can switch among clients, and each client will have its.
290 00:20:45.760 ⇒ 00:20:52.300 Samuel Roberts: Yeah, so they each… and then we added that to the URL, too, so you could theoretically, like, link to it and go back. So if I go back now, it goes back by client.
291 00:20:52.500 ⇒ 00:20:56.059 Samuel Roberts: So it’s part of the URL?
292 00:20:57.020 ⇒ 00:20:57.570 Uttam Kumaran: Okay.
293 00:20:57.830 ⇒ 00:21:01.819 Samuel Roberts: And then, yeah, you can see each one, like, this one was 4 minutes ago.
294 00:21:01.820 ⇒ 00:21:05.150 Uttam Kumaran: Cool. It has all the weekly goals, and this is the client ID up here, so…
295 00:21:05.290 ⇒ 00:21:17.339 Samuel Roberts: If… one thing I will say, I like having the client ID up here, just so, like, you can jump right to it, rather than going through a page and having to drop down. But it might get a little, like, every time you click one of these, this is a new,
296 00:21:18.860 ⇒ 00:21:22.659 Samuel Roberts: What, transition? So, like, every one of those is a back button now.
297 00:21:23.600 ⇒ 00:21:27.320 Uttam Kumaran: Excellent, I mean, I’d prefer… yeah, I don’t… I think that’s actually okay.
298 00:21:27.320 ⇒ 00:21:34.579 Samuel Roberts: That’s why I just wanted to make sure, because I could see it being where it’s all saved here, but there’s two different ways of doing it, one that adds to the history, one that doesn’t.
299 00:21:34.720 ⇒ 00:21:38.050 Samuel Roberts: I just want to make sure that that’s a good flow for you.
300 00:21:38.680 ⇒ 00:21:43.260 Samuel Roberts: And then… yeah, we kind of have a few things, like I said, the Slack stuff, we gotta…
301 00:21:43.580 ⇒ 00:21:47.040 Samuel Roberts: ingest that into Supabase.
302 00:21:47.180 ⇒ 00:21:53.070 Samuel Roberts: And the Zoom stuff, we have some ideas with Turbo Puffer, and the Heroku timeout issue, we have some ideas for too, but…
303 00:21:54.700 ⇒ 00:21:59.610 Samuel Roberts: I think that’s kind of why this is still linear heavy, is because linear is the one that’s working the best right now, so…
304 00:22:01.080 ⇒ 00:22:03.399 Samuel Roberts: But yeah, that’s kind of where we’re at.
305 00:22:04.020 ⇒ 00:22:04.870 Samuel Roberts: You know, I guess…
306 00:22:04.870 ⇒ 00:22:09.870 Uttam Kumaran: Tell me a little bit about, you know, the concept of…
307 00:22:10.410 ⇒ 00:22:16.060 Uttam Kumaran: Like, kind of the three phases of… stand-up, right? So there’s, like.
308 00:22:16.340 ⇒ 00:22:32.710 Uttam Kumaran: there’s, like, a little bit of pre-work. Like, this is actually great, because now I can actually come in here, and at least from the linear side of things, I can get a sense. Previously, see, right now, I’m just subscribed to almost every linear channel, and so I just go through my notifications.
309 00:22:33.030 ⇒ 00:22:36.189 Uttam Kumaran: Which is, like, I don’t know, it’s, like, completely useless.
310 00:22:36.320 ⇒ 00:22:42.919 Uttam Kumaran: I maybe catch, like, one or two things. So this is actually helpful, because now it’s organized. The second piece is…
311 00:22:43.350 ⇒ 00:22:49.899 Uttam Kumaran: there’s almost also, like, a retrospective thing. Like, Rico, for example, takes notes.
312 00:22:50.020 ⇒ 00:22:51.639 Samuel Roberts: During stand-up?
313 00:22:51.940 ⇒ 00:22:55.389 Uttam Kumaran: And then is able to send, like, a little update of, like.
314 00:22:55.730 ⇒ 00:22:58.919 Uttam Kumaran: What was the… what are the core priorities today?
315 00:22:58.920 ⇒ 00:22:59.850 Samuel Roberts: Yeah. Right.
316 00:22:59.960 ⇒ 00:23:06.009 Uttam Kumaran: And of course, there’s the during. So I wonder, like, what is the,
317 00:23:06.790 ⇒ 00:23:09.989 Uttam Kumaran: Like, are you guys thinking about it? Like, there is a stand-up
318 00:23:10.290 ⇒ 00:23:15.080 Uttam Kumaran: object created? Because right now, this is more of just, like.
319 00:23:15.420 ⇒ 00:23:22.599 Uttam Kumaran: it is… it is, like, a rolling, just, page with updated data. Do you see this more as, like, okay, we’re gonna save a state
320 00:23:22.750 ⇒ 00:23:25.000 Uttam Kumaran: Of, like, a stand-up object.
321 00:23:25.350 ⇒ 00:23:30.630 Uttam Kumaran: As the other piece of this also is, like, associating,
322 00:23:31.160 ⇒ 00:23:34.529 Uttam Kumaran: the video, right? So, after this is completed.
323 00:23:34.780 ⇒ 00:23:40.559 Uttam Kumaran: So I guess my, yeah, my… more of my question is, like, what are y’all thinking about, like, the pre- and post…
324 00:23:41.620 ⇒ 00:23:43.269 Uttam Kumaran: Stand up, if at all.
325 00:23:43.640 ⇒ 00:23:50.340 Samuel Roberts: So, I think we, I don’t know if I have that open in Figma still. We did talk about that. This was kind of the V1…
326 00:23:50.630 ⇒ 00:23:56.250 Samuel Roberts: was this? Let me… Y’all see the figment now?
327 00:23:56.580 ⇒ 00:23:57.040 Gabriel Lam: Yep.
328 00:23:57.040 ⇒ 00:23:57.700 Mustafa Raja: Yeah.
329 00:23:57.700 ⇒ 00:24:03.670 Samuel Roberts: So the V1 was just this stage. We did discuss this,
330 00:24:03.960 ⇒ 00:24:07.349 Samuel Roberts: Kind of idea, where there would be something like in-meeting notes.
331 00:24:07.520 ⇒ 00:24:19.220 Samuel Roberts: you know, there’d be some linear updates, and then there’d be, like, the post-meeting transcript, or whatever kind of, you know, Freaka’s putting together a succinct, like, list of priorities that could get put into a post-meeting thing. So that is…
332 00:24:19.570 ⇒ 00:24:24.300 Samuel Roberts: Not necessarily part of this, yet. We did…
333 00:24:24.420 ⇒ 00:24:34.569 Samuel Roberts: think a little bit about it as a stand-up object, and, like, I think you mentioned, like, having, like, a sprint has multiple stand-ups kind of thing,
334 00:24:35.240 ⇒ 00:24:41.280 Samuel Roberts: For now, I think the idea was to persistence in the browser, so it at least is not going away. Eventually.
335 00:24:41.660 ⇒ 00:24:51.800 Samuel Roberts: with maybe a little more data modeling, getting that into Superbase, keeping that consistent, so that it’s tracking over the days and weeks, tying those meetings to it,
336 00:24:52.130 ⇒ 00:24:54.779 Samuel Roberts: But for this kind of initial…
337 00:24:55.420 ⇒ 00:25:07.300 Samuel Roberts: initial swing, it was, let’s get all of these things ingested and into a useful kind of summary. We did talk about… yeah, there could be a place here for notes that would get added as well. I think that’s probably good.
338 00:25:07.650 ⇒ 00:25:17.689 Samuel Roberts: eventually, there could be something that ingests the transcript right after and makes sure the thing’s lined up, I suppose. That’s more to think about, I think, here, but…
339 00:25:17.880 ⇒ 00:25:23.310 Samuel Roberts: for… for V1, we were… we had… we had thought this through, but figured, let’s focus on getting this good first.
340 00:25:24.920 ⇒ 00:25:34.239 Gabriel Lam: I also feel like, just to chime in, I think for V1, we looked at it more as, like, hey, what information do I need to run a stand-up?
341 00:25:34.540 ⇒ 00:25:35.020 Samuel Roberts: Yeah.
342 00:25:35.020 ⇒ 00:25:45.040 Gabriel Lam: And then, once the stand-up is done, I think it’s a separate, you know, it’s a follow-up development to say, now that it’s done, how do we get that information back in so that tomorrow, when you see all this.
343 00:25:45.040 ⇒ 00:25:47.799 Uttam Kumaran: You know how to ruin tomorrow’s stand-up. Nice, okay.
344 00:25:48.060 ⇒ 00:25:53.069 Uttam Kumaran: Yeah, walk me through, walk me through this diagram, if you don’t mind. This is the first time I’m seeing this.
345 00:25:53.070 ⇒ 00:25:57.290 Samuel Roberts: Okay, yeah, so this was after, I think this was last week after we watched
346 00:25:57.400 ⇒ 00:26:00.940 Samuel Roberts: We had a, are…
347 00:26:00.940 ⇒ 00:26:01.740 Casie Aviles: Pretty much.
348 00:26:02.060 ⇒ 00:26:05.990 Samuel Roberts: We watched the stand-up, we all kind of took some notes here. Cool.
349 00:26:06.190 ⇒ 00:26:14.550 Uttam Kumaran: My more recent ones are a little bit rough, because we have so many clients, and I end up just focusing on, like, the biggest ones.
350 00:26:15.140 ⇒ 00:26:20.530 Samuel Roberts: No, you’re good, you’re good. Yeah, no, I mean, but this is… this is good, because, like, if we’re gonna run it this way, we need to understand that, you know?
351 00:26:20.530 ⇒ 00:26:25.730 Uttam Kumaran: No, and actually, like, I want to talk about all of them. As you can see…
352 00:26:26.130 ⇒ 00:26:31.519 Uttam Kumaran: I mean, my adjustment I probably am gonna end up making is… the next stand-up, like…
353 00:26:31.700 ⇒ 00:26:34.860 Uttam Kumaran: I feel like this, for this team, we’re gonna need this amount of time, at least within.
354 00:26:34.860 ⇒ 00:26:35.340 Samuel Roberts: Yeah.
355 00:26:35.340 ⇒ 00:26:45.019 Uttam Kumaran: Every day, right? And so, it’s the next stand-up that is gonna continue to extend. Like, I’m gonna push it again, 45 minutes, push it again, 15 minutes again, 15 minutes, because…
356 00:26:45.020 ⇒ 00:26:45.690 Samuel Roberts: Yeah.
357 00:26:45.850 ⇒ 00:26:53.980 Uttam Kumaran: there are… there actually are some clients that I do think, if I had the context, I could wrap up in, like.
358 00:26:54.620 ⇒ 00:26:59.980 Uttam Kumaran: 2 to 5 minutes. Like, honestly, probably even just 2 minutes, because I… it’s…
359 00:27:00.170 ⇒ 00:27:04.260 Uttam Kumaran: Like, I’m a… I’m a very, like, I don’t… I…
360 00:27:04.700 ⇒ 00:27:18.669 Uttam Kumaran: Stand-up is just me being like, do I need to do anything as a PM? Right. Who is on the hook for something? If we got something done, can we share it? If there’s something blocked, how do I unblock? And that’s… those are just it. And so, for every client.
361 00:27:18.800 ⇒ 00:27:26.840 Samuel Roberts: it’s like a Rolodex of, like, okay, what to ask, what to ask, what to ask, what to ask. But there are, of course, priority clients, right? So Insomnia is a priority client.
362 00:27:26.980 ⇒ 00:27:35.139 Uttam Kumaran: And they’re… and they’re, frankly, they’re most… they’re ranked in two ways. A forward revenue potential, and active revenue.
363 00:27:35.310 ⇒ 00:27:45.329 Uttam Kumaran: So Insomnia is a good situation where they are… if we nail it for them, they’ll be with us for a while. Eden is a situation of, like, they’re our biggest client. And so.
364 00:27:45.330 ⇒ 00:27:46.000 Samuel Roberts: Right.
365 00:27:46.000 ⇒ 00:27:50.119 Uttam Kumaran: no matter what happens in those stand-ups, I at least have to talk about those two.
366 00:27:50.840 ⇒ 00:27:59.720 Uttam Kumaran: And then I try to talk about the worst, the clients that we’re struggling with. And if I can get those, then the middle I’ll figure out in the middle of the day, you know, so…
367 00:28:00.010 ⇒ 00:28:00.730 Samuel Roberts: Okay.
368 00:28:01.720 ⇒ 00:28:03.720 Samuel Roberts: Yeah, I mean, I think…
369 00:28:04.620 ⇒ 00:28:14.999 Samuel Roberts: I mean, this should be very, like, you know, if we can get this to work with the right Zoom, and I mean, the Slack messages are definitely working for a client, they’re not working for internal teams, so maybe that’s less of a priority then? I mean…
370 00:28:15.000 ⇒ 00:28:20.000 Uttam Kumaran: And then for the… I guess my other question on… from my loom was on, like.
371 00:28:20.120 ⇒ 00:28:27.150 Uttam Kumaran: references, or, like, kind of retrieval, right? Like, showing… what,
372 00:28:27.330 ⇒ 00:28:35.610 Uttam Kumaran: like, I’m… I’m actually… I’m more of, like, this is really, really strong, like, if I can literally see.
373 00:28:35.810 ⇒ 00:28:38.940 Samuel Roberts: the relevant Slack messages that formed.
374 00:28:39.520 ⇒ 00:28:44.440 Uttam Kumaran: I mean, I mean, it, it, to be, to be… again, I… this is, like.
375 00:28:44.960 ⇒ 00:28:48.849 Uttam Kumaran: this is, like, if I was to record another loom here, this is kind of what I would think about, is…
376 00:28:48.850 ⇒ 00:28:49.240 Samuel Roberts: Yeah, yeah.
377 00:28:49.240 ⇒ 00:28:53.330 Uttam Kumaran: For example, can you scroll down here?
378 00:28:55.330 ⇒ 00:29:02.300 Uttam Kumaran: let’s say… let’s say there is an item here from Slack. Are… are you planning on linking, like,
379 00:29:02.830 ⇒ 00:29:06.350 Uttam Kumaran: like, basically permalinking to the message? Yeah. Okay.
380 00:29:06.350 ⇒ 00:29:24.810 Samuel Roberts: Yeah, that’s like, I think, you know, we were having issues where we weren’t getting things even showing up from Slack to begin with. Linear was kind of dominated in context, and that one’s easy because they all have IDs, and it’s all being pulled straight from Linear, so you can see the movement. Slack, we’re pulling from the Superbase, which is pulling from Slack, so there was a little more work to do to, like.
381 00:29:24.810 ⇒ 00:29:25.839 Uttam Kumaran: Okay.
382 00:29:26.180 ⇒ 00:29:31.899 Samuel Roberts: I refloat it a little bit for the context, so I wasn’t just passing in, like, a giant JSON object.
383 00:29:31.900 ⇒ 00:29:46.670 Samuel Roberts: So I think that’s where we lose a little bit of the reference, but we can clean that up, and I want to clean that up. This was kind of the first pass, was, let’s just get conversations from Slack into there. I think what we need to do is add the message IDs, because that would definitely be able to very easily link here, I think.
384 00:29:47.240 ⇒ 00:30:06.290 Uttam Kumaran: And there’s kind of, like, two things I would say that’s helpful here. One is, if you can at least isolate all the relevant Slack messages from yesterday, as you can tell in the meetings, I’m solving that by just… I just pull up Slack and go to the channel, and, like, scroll up, right? So, at minimum, even if the AI isn’t perfect.
385 00:30:06.470 ⇒ 00:30:10.859 Uttam Kumaran: as long as all the references are here, I would rather just click on your guys’.
386 00:30:10.860 ⇒ 00:30:12.800 Samuel Roberts: Poll, and just…
387 00:30:12.800 ⇒ 00:30:19.770 Uttam Kumaran: it would see… I would… I could see, like, okay, here are all the Slack messages, tell me which one is relevant, right? That is a step in the right direction.
388 00:30:19.770 ⇒ 00:30:22.189 Samuel Roberts: Of course, the last mile step here is, like.
389 00:30:22.280 ⇒ 00:30:23.750 Uttam Kumaran: Take that, and then…
390 00:30:23.950 ⇒ 00:30:27.979 Uttam Kumaran: Take the 29 and show me the 3 that are relevant for one of these, right?
391 00:30:27.980 ⇒ 00:30:29.330 Samuel Roberts: Right, right, right.
392 00:30:30.310 ⇒ 00:30:33.069 Samuel Roberts: Okay, yeah, yeah, I mean, that’s definitely what we’re working towards. I think it’s.
393 00:30:33.070 ⇒ 00:30:41.780 Uttam Kumaran: What is the… and so the mechanics right now is we’re using Polyatomic to sync this to Snowflake, and then you’re moving… like, how is this working?
394 00:30:41.780 ⇒ 00:30:47.870 Samuel Roberts: So right now, the Slack messages are already synced. I don’t remember exactly what is doing that. Casey was stopping, you might…
395 00:30:47.870 ⇒ 00:30:51.020 Mustafa Raja: I think Daxter, we’re using Daxter for that.
396 00:30:51.020 ⇒ 00:30:52.130 Casie Aviles: Yeah, Dexter.
397 00:30:52.130 ⇒ 00:30:54.490 Samuel Roberts: So yeah, like, if I go to,
398 00:30:55.870 ⇒ 00:30:58.310 Samuel Roberts: Shh, this is the wrong one, gotta go to…
399 00:30:58.930 ⇒ 00:31:03.250 Samuel Roberts: Slack. Slack. So there’s a whole project here that’s got all the…
400 00:31:04.080 ⇒ 00:31:08.179 Samuel Roberts: Messages, and the embeddings, and this is how we’re doing, like, vector search over them.
401 00:31:08.180 ⇒ 00:31:11.079 Uttam Kumaran: And then, I guess, tell me about how the Daxter pipeline works.
402 00:31:12.930 ⇒ 00:31:16.889 Casie Aviles: Yeah, for the dog start pipeline,
403 00:31:17.190 ⇒ 00:31:20.509 Casie Aviles: What it’s doing, basically, is it gets, like, the raw…
404 00:31:21.730 ⇒ 00:31:25.679 Casie Aviles: It gets the raw export from Polyatomic, from S3.
405 00:31:26.280 ⇒ 00:31:31.259 Casie Aviles: And then it transforms that into client tables.
406 00:31:31.770 ⇒ 00:31:37.269 Casie Aviles: So, it gets, perfect. And it’s only the channels where the bot is there.
407 00:31:37.880 ⇒ 00:31:39.380 Casie Aviles: Yes, yes, exactly.
408 00:31:39.610 ⇒ 00:31:45.180 Samuel Roberts: That’s why we realized earlier that we’re missing it on the, like, AI. This is… this is really zero. It’s just…
409 00:31:45.180 ⇒ 00:31:50.540 Uttam Kumaran: Yes. And then, I guess, give me a sense, Casey, when the bot is there, we can’t go backwards?
410 00:31:52.480 ⇒ 00:31:58.699 Casie Aviles: I haven’t… yeah, I haven’t thoroughly checked, but I believe it should be…
411 00:31:58.700 ⇒ 00:32:00.789 Samuel Roberts: Backwards from when the ball was installed, you mean?
412 00:32:01.570 ⇒ 00:32:02.120 Casie Aviles: Yeah.
413 00:32:04.560 ⇒ 00:32:05.629 Samuel Roberts: Yeah, that I don’t know.
414 00:32:05.630 ⇒ 00:32:07.690 Casie Aviles: Double check. I have to double check.
415 00:32:07.690 ⇒ 00:32:16.590 Uttam Kumaran: So, I think, like, in this situation, there’s some clients where when we initialize the Slack channel, we add the bot. I would like us to…
416 00:32:17.600 ⇒ 00:32:22.209 Uttam Kumaran: If you confirm that with me, I can add it to the other channels. I would like to inform…
417 00:32:23.020 ⇒ 00:32:26.079 Uttam Kumaran: The clients, like, what it is it’s for.
418 00:32:26.240 ⇒ 00:32:34.179 Uttam Kumaran: basically, I’m gonna say this is just helping us keep track of important Black messages.
419 00:32:34.600 ⇒ 00:32:36.990 Samuel Roberts: Oh, yeah, because this is all the internal channels, right?
420 00:32:36.990 ⇒ 00:32:44.670 Uttam Kumaran: Well, a lot… yeah, a lot of these you’ll see as client, you won’t see the external… yeah, you may… you may not see them, I don’t know.
421 00:32:44.880 ⇒ 00:32:48.830 Uttam Kumaran: So… Yeah, this is…
422 00:32:49.280 ⇒ 00:32:50.130 Samuel Roberts: Alright.
423 00:32:50.730 ⇒ 00:32:56.349 Uttam Kumaran: I think it just, I would just do an audit of, like, what channels you have access to.
424 00:32:56.480 ⇒ 00:32:59.849 Uttam Kumaran: Several of these clients, of course, are now, like.
425 00:33:00.730 ⇒ 00:33:03.570 Samuel Roberts: I’m sure there’s nothing… there’s nothing here from recent.
426 00:33:03.570 ⇒ 00:33:09.749 Uttam Kumaran: Yeah, but I guess what’s… there’s a lot… we have a lot of external client… I mean, all of our client channels are.
427 00:33:09.750 ⇒ 00:33:13.230 Samuel Roberts: Divide in two ways, right? Our internal, external, so…
428 00:33:13.510 ⇒ 00:33:19.120 Uttam Kumaran: You let me know. I can… there are… I can test it, like, on Hype, for example.
429 00:33:20.130 ⇒ 00:33:24.250 Uttam Kumaran: But yeah, just let me know, like, when you guys get to that step.
430 00:33:25.060 ⇒ 00:33:30.290 Samuel Roberts: Okay, yeah, I mean, we’re currently, right now, pulling from NIST, which I think is… pulling…
431 00:33:30.450 ⇒ 00:33:34.040 Samuel Roberts: New things from the internal client channels.
432 00:33:34.480 ⇒ 00:33:39.050 Samuel Roberts: So, if we wanted to add the external ones.
433 00:33:40.880 ⇒ 00:33:47.279 Samuel Roberts: We would need to run that a new… add the bot to those channels, and then create new tables, is that right?
434 00:33:47.800 ⇒ 00:33:49.350 Samuel Roberts: Casey, you stop on?
435 00:33:50.120 ⇒ 00:33:50.720 Mustafa Raja: Yep.
436 00:33:51.360 ⇒ 00:34:00.900 Samuel Roberts: The other side of this is, and this is something that I’ve been debating a little bit with the whole moving client hubs into code and generalizing it, is are separate tables
437 00:34:01.020 ⇒ 00:34:06.620 Samuel Roberts: The way we want to go, or should this have a column that is, like, client ID.
438 00:34:06.790 ⇒ 00:34:10.310 Samuel Roberts: Because right now, we have to create these every time. Yeah.
439 00:34:10.310 ⇒ 00:34:11.979 Uttam Kumaran: I guess I would.
440 00:34:11.980 ⇒ 00:34:14.469 Samuel Roberts: I’m wondering, yeah, what the thought there is on…
441 00:34:16.780 ⇒ 00:34:21.709 Samuel Roberts: Keeping client data separated versus having to query it separately.
442 00:34:21.719 ⇒ 00:34:25.149 Uttam Kumaran: It’s gonna be a future issue in terms of security.
443 00:34:25.150 ⇒ 00:34:25.880 Samuel Roberts: That’s what, yeah.
444 00:34:25.889 ⇒ 00:34:28.649 Uttam Kumaran: one or two ways. You can continue this way.
445 00:34:29.689 ⇒ 00:34:33.479 Uttam Kumaran: Or, you can do it the other way.
446 00:34:43.849 ⇒ 00:34:46.069 Uttam Kumaran: I guess I don’t have a strong opinion.
447 00:34:46.500 ⇒ 00:34:47.899 Samuel Roberts: That’s kind of where I settled in.
448 00:34:48.209 ⇒ 00:34:49.289 Uttam Kumaran: Awake. Yeah.
449 00:34:49.460 ⇒ 00:34:50.130 Samuel Roberts: Okay.
450 00:34:50.420 ⇒ 00:34:51.879 Uttam Kumaran: What he thinks.
451 00:34:51.889 ⇒ 00:35:00.619 Samuel Roberts: Because I understand the idea of keeping them separate, but I also understand the idea of, like, we could keep them separate, we could do some RLS rules that, like, you know, they can only query the right.
452 00:35:01.419 ⇒ 00:35:03.239 Mustafa Raja: Entries and things like that.
453 00:35:03.239 ⇒ 00:35:09.499 Samuel Roberts: Either way, it’s, you know, it’s gonna add some complexity either way, because either we need to try to figure out how to programmatically create these.
454 00:35:10.419 ⇒ 00:35:15.149 Samuel Roberts: Or we need to figure out the good RLS rules to keep them Separated.
455 00:35:17.559 ⇒ 00:35:27.159 Samuel Roberts: I will… I will message your wife about that, though. It’s a good idea. Because, yeah, as I was trying to think if there are, like, there’s pros and cons to each one, and I’m definitely thinking more on the, like, making it more…
456 00:35:27.369 ⇒ 00:35:29.949 Samuel Roberts: You know, making a generic client hub.
457 00:35:29.950 ⇒ 00:35:30.340 Uttam Kumaran: Yeah.
458 00:35:30.740 ⇒ 00:35:35.949 Uttam Kumaran: go wish what he thinks, like, I think you’ll… I… I… you can also probably just…
459 00:35:36.060 ⇒ 00:35:38.470 Uttam Kumaran: Take all these and shove it into something?
460 00:35:41.410 ⇒ 00:35:43.839 Uttam Kumaran: Yeah, I don’t know, ask him what he thinks.
461 00:35:44.220 ⇒ 00:35:52.690 Samuel Roberts: Okay, because, I mean, like, this is all… this is… this is just client messages, or Slack messages for clients. We’re not necessarily storing anything that we’re not already just chatting about on,
462 00:35:54.000 ⇒ 00:35:58.739 Samuel Roberts: on Slack, so… I suppose if there’s real, like…
463 00:35:59.050 ⇒ 00:36:06.550 Samuel Roberts: data issues, like, is Slack… even though… you know what I mean? Like, there’s already… Slack is already storing it, so I don’t know, but…
464 00:36:06.680 ⇒ 00:36:08.330 Samuel Roberts: Okay.
465 00:36:08.710 ⇒ 00:36:12.520 Samuel Roberts: We can message away about that, but that was something I was already thinking about for that, and for even the,
466 00:36:12.900 ⇒ 00:36:17.040 Samuel Roberts: Meetings, because these are all… also separated.
467 00:36:17.270 ⇒ 00:36:18.350 Uttam Kumaran: I see, okay.
468 00:36:18.350 ⇒ 00:36:28.050 Samuel Roberts: And as I’m thinking about putting, you know, putting this into code and making it so that instead of having, you know, several different N8N flows, we have one master flow that just takes a client as an input.
469 00:36:28.300 ⇒ 00:36:32.650 Samuel Roberts: And then it doesn’t have to be created every time.
470 00:36:32.830 ⇒ 00:36:34.170 Samuel Roberts: Let me get a new client.
471 00:36:35.230 ⇒ 00:36:38.469 Uttam Kumaran: Yeah, it’s probably worth… worth asking him, I…
472 00:36:38.470 ⇒ 00:36:39.160 Samuel Roberts: Yeah, definitely.
473 00:36:39.160 ⇒ 00:36:39.840 Uttam Kumaran: Yeah.
474 00:36:40.040 ⇒ 00:36:40.980 Samuel Roberts: That’s a good call.
475 00:36:41.330 ⇒ 00:36:42.500 Samuel Roberts: Okay, cool.
476 00:36:43.030 ⇒ 00:36:44.150 Samuel Roberts: That’s helpful.
477 00:36:45.460 ⇒ 00:36:53.050 Uttam Kumaran: Okay, so what’s, I guess tell me… yeah, maybe tell me a little bit more about the plan for today, and then kind of, like, how we’re driving
478 00:36:53.250 ⇒ 00:36:54.740 Uttam Kumaran: towards Friday.
479 00:36:54.740 ⇒ 00:37:03.469 Samuel Roberts: Yeah, so, what we were discussing was the internal Doesn’t have any slack, which… We can…
480 00:37:05.500 ⇒ 00:37:10.960 Samuel Roberts: figure out how to get those in there. Maybe it’s just the bot. The… this Zoom…
481 00:37:11.350 ⇒ 00:37:17.420 Samuel Roberts: Issue is the fact that, like, these things are not relevant to this
482 00:37:17.700 ⇒ 00:37:26.189 Samuel Roberts: And so what we’re thinking about is Turbo Puffer to help us find the right meetings, because there’s a few that overlap, and then,
483 00:37:27.450 ⇒ 00:37:34.840 Samuel Roberts: Adding another step to pre… Pre-select the parts of the meeting that are relevant to that client.
484 00:37:35.110 ⇒ 00:37:37.720 Uttam Kumaran: Okay. Since it’s all one big meeting now.
485 00:37:37.810 ⇒ 00:37:42.599 Samuel Roberts: And it… I think we’re just kind of overloading the context with a giant transcript of…
486 00:37:43.600 ⇒ 00:37:49.800 Samuel Roberts: Multiple clients, so let’s pre-process that, and then pass that into the general summary agent.
487 00:37:51.610 ⇒ 00:38:02.180 Samuel Roberts: The other thing was the Heroku timeout that we ran into. We probably want to change it so that when we hit this button, we’re streaming back a little bit more information, because right now, Heroku times out at
488 00:38:02.400 ⇒ 00:38:04.309 Samuel Roberts: 30 seconds, you said, Ms. Dappa?
489 00:38:04.320 ⇒ 00:38:08.370 Uttam Kumaran: What is… what is the… what is it, like, what’s the prompt?
490 00:38:09.790 ⇒ 00:38:13.979 Samuel Roberts: I think you… Did you change the models, Mustafa, to figure it out?
491 00:38:13.980 ⇒ 00:38:31.239 Mustafa Raja: Yes, so, I changed the model to O4 Mini, and O4 Mini takes more than 30 seconds to, give us the response. And what Heroku does is it only listens to an API call for 30 seconds. If it takes more than that, it just closes the,
492 00:38:31.240 ⇒ 00:38:32.639 Uttam Kumaran: Is that, like, a…
493 00:38:33.120 ⇒ 00:38:33.940 Mustafa Raja: Heroquifu.
494 00:38:34.250 ⇒ 00:38:35.840 Samuel Roberts: It’s a, it’s a, yeah.
495 00:38:35.840 ⇒ 00:38:38.319 Uttam Kumaran: Is that your issue, or, like, we have to pay for something?
496 00:38:38.320 ⇒ 00:38:40.730 Samuel Roberts: No, no, no, I think it’s just there… if a…
497 00:38:41.140 ⇒ 00:38:43.560 Uttam Kumaran: Oh, they should do it as, like, an async thing.
498 00:38:43.560 ⇒ 00:38:47.700 Samuel Roberts: Yeah, if it hits an endpoint, and the endpoint doesn’t respond for 30 seconds, it probably assumes it’s…
499 00:38:47.830 ⇒ 00:38:52.240 Samuel Roberts: I see. Messed up. So I think what we need to do is just have a small thing that just is, like.
500 00:38:52.330 ⇒ 00:39:09.080 Samuel Roberts: pulling Zoom, pulling Slack, pulling Linear, and just responding, kind of, streaming it back a little bit, and then I think it’ll let us run much longer. So that… that… again, it’s not an issue here, but as we’re gonna start pulling from Turbo Puffer, then back to Superbase to get the right meetings and everything, it’s gonna add a little more… a little more time there.
501 00:39:09.140 ⇒ 00:39:13.560 Samuel Roberts: And we just want to make sure it’s not timing out. So that was the other thing we discussed.
502 00:39:13.670 ⇒ 00:39:19.250 Samuel Roberts: But I think the biggest thing is, especially for, like, clients, if the SOC messages are probably pretty good right now, it’s.
503 00:39:19.250 ⇒ 00:39:21.680 Uttam Kumaran: So let’s say, let’s… let’s go to,
504 00:39:22.810 ⇒ 00:39:25.470 Uttam Kumaran: Let’s go to, like, Eden, for example.
505 00:39:25.470 ⇒ 00:39:26.060 Samuel Roberts: Sure.
506 00:39:36.580 ⇒ 00:39:39.270 Uttam Kumaran: Okay, so let me just, like, kind of live…
507 00:39:39.760 ⇒ 00:39:40.410 Samuel Roberts: Yeah, totally.
508 00:39:40.410 ⇒ 00:39:42.189 Uttam Kumaran: blog, what I…
509 00:39:43.010 ⇒ 00:39:50.210 Uttam Kumaran: So, today, what I would do is I would go ahead and copy back my weekly goals. The second thing is…
510 00:39:51.170 ⇒ 00:40:00.830 Uttam Kumaran: I… again, I like that it says close to urgent tickets. What I don’t like is I don’t know what those… I have to click into the ticket to see the titles and, like, what they consist of, right?
511 00:40:00.830 ⇒ 00:40:01.469 Samuel Roberts: Yeah, what is…
512 00:40:01.470 ⇒ 00:40:02.530 Uttam Kumaran: on yesterday.
513 00:40:03.660 ⇒ 00:40:06.760 Samuel Roberts: Oh, here, close 200 tickets. Oh, I see. Yeah, we can…
514 00:40:07.320 ⇒ 00:40:08.450 Uttam Kumaran: So you’ll just have to, like.
515 00:40:08.450 ⇒ 00:40:09.250 Samuel Roberts: Sweet that.
516 00:40:09.550 ⇒ 00:40:12.339 Uttam Kumaran: I would say go… go more verbose.
517 00:40:12.710 ⇒ 00:40:13.360 Samuel Roberts: Yeah.
518 00:40:13.850 ⇒ 00:40:15.329 Uttam Kumaran: Oh, way more verbose.
519 00:40:15.330 ⇒ 00:40:15.990 Mustafa Raja: Yeah, yeah.
520 00:40:15.990 ⇒ 00:40:19.769 Uttam Kumaran: I don’t care about the information density at this point. We’re not going, like…
521 00:40:19.930 ⇒ 00:40:22.700 Uttam Kumaran: I would say it’s already, like, looking great.
522 00:40:22.960 ⇒ 00:40:24.230 Uttam Kumaran: I would just, like…
523 00:40:24.570 ⇒ 00:40:29.260 Uttam Kumaran: What… what you’re trying to compete with is… is the linear view of, like, every, like.
524 00:40:29.980 ⇒ 00:40:34.300 Mustafa Raja: what I usually do, right? So go verb… go more verbose, don’t worry about…
525 00:40:34.780 ⇒ 00:40:41.180 Uttam Kumaran: Like, compressing updates. Tell the AI to opt for…
526 00:40:41.480 ⇒ 00:40:47.400 Uttam Kumaran: giving everything, because I don’t care… my… I’m optimizing not for…
527 00:40:47.710 ⇒ 00:40:50.239 Uttam Kumaran: I’m optimizing that everything’s in one place.
528 00:40:50.240 ⇒ 00:40:50.830 Samuel Roberts: Yeah.
529 00:40:51.530 ⇒ 00:40:56.499 Uttam Kumaran: first, you know, versus, like, is it the densest thing?
530 00:40:57.500 ⇒ 00:41:00.730 Uttam Kumaran: So, okay, today there’s this urgent item.
531 00:41:01.900 ⇒ 00:41:07.849 Uttam Kumaran: And it looks like there’s 16 linear issues. My other question overall is, there’s 16 linear issues.
532 00:41:08.400 ⇒ 00:41:16.980 Uttam Kumaran: That are relevant. I think it is still helpful for me to see how many linear issues, like, kind of get a sense of the cycle.
533 00:41:17.640 ⇒ 00:41:20.530 Uttam Kumaran: The cycle is…
534 00:41:20.790 ⇒ 00:41:25.699 Uttam Kumaran: attached to what this stand-up is attached to, right? So it’s also helpful for me to see, like.
535 00:41:25.870 ⇒ 00:41:27.700 Uttam Kumaran: What is in the backlog?
536 00:41:30.420 ⇒ 00:41:37.700 Uttam Kumaran: Right, and again, that doesn’t have to be a list, but it can be similarly in this… in somewhere at the bottom, like, what’s coming up this week.
537 00:41:37.950 ⇒ 00:41:39.019 Samuel Roberts: Sure.
538 00:41:39.020 ⇒ 00:41:43.550 Uttam Kumaran: I actually think this is… this is… Helpful.
539 00:41:44.100 ⇒ 00:41:46.449 Uttam Kumaran: what I will probably do today…
540 00:41:47.180 ⇒ 00:41:49.199 Uttam Kumaran: is I can run it with this.
541 00:41:49.960 ⇒ 00:41:52.779 Uttam Kumaran: if this… yeah, if I can access that, I can run it with this.
542 00:41:52.800 ⇒ 00:41:54.140 Samuel Roberts: Yeah. And then…
543 00:41:58.100 ⇒ 00:41:58.970 Uttam Kumaran: sort of…
544 00:41:59.180 ⇒ 00:42:06.139 Uttam Kumaran: ask the team, is there anything here that, like, we need… what else we need to talk about? You mind going to… yeah, you mind going to,
545 00:42:07.790 ⇒ 00:42:09.190 Uttam Kumaran: Insomnia?
546 00:42:09.190 ⇒ 00:42:10.840 Samuel Roberts: Sure.
547 00:42:22.520 ⇒ 00:42:26.050 Uttam Kumaran: Nice, guys, this is good. Oh, this is so nice.
548 00:42:26.050 ⇒ 00:42:28.500 Samuel Roberts: Oh yeah, so there’s way more Slack stuff here, dude.
549 00:42:28.650 ⇒ 00:42:29.710 Uttam Kumaran: Great, wow.
550 00:42:29.710 ⇒ 00:42:32.760 Samuel Roberts: I’d say it’s still pulling 3 Zoom stand-ups, so it’s probably overloading the context a little.
551 00:42:32.760 ⇒ 00:42:37.939 Uttam Kumaran: Yeah, and then another piece of feedback, I’m pretty sure… click on INS211, like Demolades.
552 00:42:37.940 ⇒ 00:42:39.159 Samuel Roberts: This one here?
553 00:42:39.160 ⇒ 00:42:44.910 Uttam Kumaran: onto that, yeah I think he just… Moved it to PR review.
554 00:42:44.910 ⇒ 00:42:45.910 Samuel Roberts: 52 minutes, I guess.
555 00:42:45.910 ⇒ 00:42:46.939 Uttam Kumaran: Two minutes ago?
556 00:42:48.130 ⇒ 00:42:51.850 Uttam Kumaran: Okay, so it doesn’t say, like, exactly the status of it?
557 00:42:52.260 ⇒ 00:42:58.159 Uttam Kumaran: But, again, like, again, this is where, like, in your, in your summary of the ticket, or tickets.
558 00:42:58.450 ⇒ 00:43:13.170 Uttam Kumaran: toss everything I may need to know. What is… what do I… what are the dimensionalities I need to know about a ticket? I need to know who it was assigned to, when it was created, when was the last update, what is it, how does it relate to our goals, right? Like…
559 00:43:13.480 ⇒ 00:43:26.230 Uttam Kumaran: I don’t care if there’s a lot there, because I can read it, like… I’ll read it fast, and my question is not related to, like, what I’m reading. More of the problem is that I can’t see all of those pieces in one place right now.
560 00:43:26.550 ⇒ 00:43:34.139 Uttam Kumaran: Right, so… Structure in a way that makes sense, but go for… be verbose,
561 00:43:35.340 ⇒ 00:43:37.510 Uttam Kumaran: Okay, I mean, these all look…
562 00:43:38.910 ⇒ 00:43:48.830 Uttam Kumaran: Pretty good. A couple of other things that I would, also… I’m gonna… I’m just kind of keeping a little bit of a running list, is add area for…
563 00:43:49.130 ⇒ 00:43:51.450 Uttam Kumaran: Sort of stand-up notes.
564 00:43:51.450 ⇒ 00:43:52.429 Samuel Roberts: That’s… yeah, that’s the…
565 00:43:52.430 ⇒ 00:43:55.730 Uttam Kumaran: One thing that I think you’ll be able to… you want…
566 00:43:56.060 ⇒ 00:44:04.009 Uttam Kumaran: like, I think also having a free text area for me to write, like, because Rico is writing stuff in Notion.
567 00:44:04.580 ⇒ 00:44:09.510 Uttam Kumaran: Related to what we talk about. We could easily say, just go ahead and write it in here.
568 00:44:09.860 ⇒ 00:44:14.429 Uttam Kumaran: And it will get leveraged as additional context for tomorrow.
569 00:44:14.800 ⇒ 00:44:17.450 Uttam Kumaran: Right? Or for the post-stand-up summary.
570 00:44:17.830 ⇒ 00:44:22.559 Uttam Kumaran: So that’s where it’s like, if you think about it, and I think…
571 00:44:22.890 ⇒ 00:44:25.259 Uttam Kumaran: Gabe, this is sort of how I think about…
572 00:44:25.960 ⇒ 00:44:43.990 Uttam Kumaran: the AI, like, kind of, like, context problem is you have a mix of structured data, right? Zooms, transcripts, Slack messages, linear. You also are going to have, like, free text. And the free text is probably going to be with the distances between 80% to 100%.
573 00:44:44.170 ⇒ 00:44:49.499 Uttam Kumaran: Because that’s where we’re gonna… if it makes it into, like, a… like, something I need to write down.
574 00:44:49.500 ⇒ 00:44:50.080 Samuel Roberts: Right.
575 00:44:50.080 ⇒ 00:45:02.800 Uttam Kumaran: probably something that… it could be something that isn’t… hasn’t been said out loud, or I was like, hey, okay, I’ll… like, I’m not usually, like… it may be just be something that could be helpful, so having a free text area to write
576 00:45:03.100 ⇒ 00:45:04.590 Uttam Kumaran: Stand-up updates?
577 00:45:04.980 ⇒ 00:45:08.260 Uttam Kumaran: That then get leveraged For the next…
578 00:45:09.210 ⇒ 00:45:12.300 Uttam Kumaran: Day stand-up could be helpful, or it could be, again, like,
579 00:45:12.420 ⇒ 00:45:20.399 Uttam Kumaran: Again, if you don’t have the concept of a stand-up object, at least what you can do is… and there’s a different… again, we can think about the designs as we go, but…
580 00:45:21.470 ⇒ 00:45:28.350 Uttam Kumaran: At least what could be helpful is for each stand-up to have an ability to, like, be like, today’s updates, today’s updates, today’s updates.
581 00:45:28.350 ⇒ 00:45:30.230 Samuel Roberts: Yeah, yeah. I mean, we can…
582 00:45:30.530 ⇒ 00:45:31.140 Uttam Kumaran: Yeah.
583 00:45:31.350 ⇒ 00:45:33.219 Uttam Kumaran: Sorry. I, I, I…
584 00:45:33.640 ⇒ 00:45:47.280 Uttam Kumaran: Yeah, that’s a… that’s a broader design issue. Like, I could see you guys creating an entire sprint object, creating a project object. Within a project, you create a sprint object. Within Sprint, there’s, like, stand-about… I don’t want to, like, think too hard about it, but…
585 00:45:47.900 ⇒ 00:45:50.179 Uttam Kumaran: If there’s an ability to put in free text.
586 00:45:50.290 ⇒ 00:45:53.580 Uttam Kumaran: we can move Rico’s workflow here as well.
587 00:45:53.580 ⇒ 00:45:55.529 Samuel Roberts: Yeah. Yeah, we could definitely…
588 00:45:55.680 ⇒ 00:46:01.550 Samuel Roberts: add that… I think the question becomes, like, yeah, where does that get stored, and do we want to start doing that kind of, like.
589 00:46:01.990 ⇒ 00:46:05.910 Samuel Roberts: Additional, like, logging these into a data… a standard.
590 00:46:05.910 ⇒ 00:46:10.210 Uttam Kumaran: I would, I mean, what’s the… like, I would say, what’s… yeah, I’d say, what’s the cost?
591 00:46:10.210 ⇒ 00:46:15.589 Samuel Roberts: I was just, like, let’s just… we were just trying to focus on getting this to be useful. Oh, sure, sure, sure, sure, sure. But if the free text is…
592 00:46:15.590 ⇒ 00:46:18.029 Uttam Kumaran: If you can at least give me a free text box…
593 00:46:18.030 ⇒ 00:46:18.670 Samuel Roberts: Yeah, we can throw it.
594 00:46:18.670 ⇒ 00:46:20.350 Uttam Kumaran: persist in my browser?
595 00:46:20.600 ⇒ 00:46:22.119 Samuel Roberts: Yeah, that’ll be easy enough to do.
596 00:46:22.740 ⇒ 00:46:36.829 Samuel Roberts: And that could even… the question becomes, like, we’ll just add that to the context when you click refresh for now, but eventually that would go and leapfrog day-to-day, previous day’s context. But I think, yeah, for now, I would say,
597 00:46:37.270 ⇒ 00:46:40.369 Samuel Roberts: We were just trying to, like, nail this before we start worrying about, like.
598 00:46:40.370 ⇒ 00:46:41.110 Uttam Kumaran: Sure, sure, sure.
599 00:46:41.110 ⇒ 00:46:47.799 Samuel Roberts: multiple weeks, and there’s still plenty to do there. But yeah, we can add a free text that’s at least part of the context. Honestly, like, I think in the…
600 00:46:48.210 ⇒ 00:46:50.810 Samuel Roberts: Like, is this… is multiple of these…
601 00:46:51.530 ⇒ 00:46:53.979 Samuel Roberts: Better, or would you want a free text for the goals, too?
602 00:46:53.980 ⇒ 00:46:55.489 Uttam Kumaran: No, that’s okay.
603 00:46:55.490 ⇒ 00:46:56.080 Samuel Roberts: Okay.
604 00:46:56.850 ⇒ 00:47:00.319 Samuel Roberts: We could have weekly goals and notes, maybe, and then all the stuff.
605 00:47:00.700 ⇒ 00:47:01.829 Samuel Roberts: Does that work?
606 00:47:01.830 ⇒ 00:47:02.819 Uttam Kumaran: Yeah, that’s fine.
607 00:47:02.820 ⇒ 00:47:07.809 Samuel Roberts: Okay, and then, yeah, let’s persistent, because I think… I think you’re absolutely right. We want to get to that point where we can feed in each day.
608 00:47:07.810 ⇒ 00:47:11.050 Uttam Kumaran: Because the Notion shit we’re writing is getting lost.
609 00:47:11.050 ⇒ 00:47:11.730 Samuel Roberts: Yeah.
610 00:47:11.730 ⇒ 00:47:15.250 Uttam Kumaran: And I want to point people to just, like, go check out our stand-up.
611 00:47:15.370 ⇒ 00:47:16.440 Samuel Roberts: Object.
612 00:47:16.440 ⇒ 00:47:17.450 Uttam Kumaran: In the platform, right?
613 00:47:17.450 ⇒ 00:47:23.190 Samuel Roberts: Yeah, no, I think… I think we’re getting to that point where that’s definitely what this would… this will become.
614 00:47:23.320 ⇒ 00:47:24.779 Samuel Roberts: But it’s just a lot of, like.
615 00:47:25.210 ⇒ 00:47:25.660 Uttam Kumaran: Yeah.
616 00:47:25.660 ⇒ 00:47:30.430 Samuel Roberts: Figure out all that storage that’s gonna distract from, like, Tuning this properly.
617 00:47:31.420 ⇒ 00:47:33.420 Uttam Kumaran: A couple of…
618 00:47:34.410 ⇒ 00:47:44.210 Uttam Kumaran: messages that I also send in, maybe, Gabe, I can get… give you the last word before we transition. One is, like, we should think about some way for users to submit bugs easily.
619 00:47:44.650 ⇒ 00:47:47.100 Uttam Kumaran: Across the… any… across the platform in general.
620 00:47:47.100 ⇒ 00:47:47.840 Samuel Roberts: Yes.
621 00:47:47.840 ⇒ 00:47:55.309 Uttam Kumaran: know right now, I screenshot, and I send something, and I send it to Slack, and then I create a linear. I know there’s some helpful things.
622 00:47:55.310 ⇒ 00:47:58.339 Samuel Roberts: We need a little thing that goes down the corner that’s just like, here’s a bug.
623 00:47:58.340 ⇒ 00:48:06.409 Uttam Kumaran: Yeah, right, like, I don’t know, what is it, like, I feel like, is it Sentry, or… I don’t… or not Sentry, there’s a… there’s some things that, like, are helpful for that, but…
624 00:48:06.410 ⇒ 00:48:06.880 Samuel Roberts: Yeah.
625 00:48:06.880 ⇒ 00:48:09.880 Uttam Kumaran: I’m… again, now I’m kind of going broader than me. Second.
626 00:48:10.260 ⇒ 00:48:15.020 Uttam Kumaran: They’re… just put tooltips, like, in as many places as possible.
627 00:48:15.450 ⇒ 00:48:15.849 Samuel Roberts: Not true.
628 00:48:15.850 ⇒ 00:48:22.470 Uttam Kumaran: So that it’s just helpful, because all these questions I have is, like, how did you pick these?
629 00:48:23.020 ⇒ 00:48:33.140 Uttam Kumaran: what, like, what… what does changes mean? So these are all, like, late-stage problems. Third is, you should do one-on-one testing with Amber and Rico first.
630 00:48:33.690 ⇒ 00:48:38.020 Uttam Kumaran: Once you get… once this is to, like, a place closer, probably, to Friday.
631 00:48:38.230 ⇒ 00:48:42.579 Uttam Kumaran: When you’re refreshing this, don’t, block the screen.
632 00:48:42.770 ⇒ 00:48:47.970 Uttam Kumaran: Because I may just hit refresh, and then keep, like… Keep going.
633 00:48:47.970 ⇒ 00:48:50.180 Samuel Roberts: For sure, yeah, we’re gonna have the spinner there instead of the… yeah, that’s smart.
634 00:48:50.180 ⇒ 00:48:53.379 Uttam Kumaran: Yeah. And then, yeah, I feel like that’s it.
635 00:48:53.380 ⇒ 00:48:54.010 Gabriel Lam: Okay.
636 00:48:54.450 ⇒ 00:48:57.779 Uttam Kumaran: Yeah, I guess, Gabe, I’ll give you the last word on anything else.
637 00:48:58.690 ⇒ 00:49:06.379 Gabriel Lam: I… I think we covered pretty much everything. I had put in a little,
638 00:49:07.770 ⇒ 00:49:15.570 Gabriel Lam: set of core features that… or additional features to think about for Friday, mostly if there’s ways to have inline controls to…
639 00:49:15.970 ⇒ 00:49:25.179 Gabriel Lam: close our ad blockers from the screen, but I think at the moment, we saw it as, like, a, hey, how do we run a stand-up as opposed to
640 00:49:25.580 ⇒ 00:49:32.560 Gabriel Lam: like, how do we add, you know, specific links to linear? I think that’s a subsequent feature that…
641 00:49:32.700 ⇒ 00:49:47.069 Gabriel Lam: is maybe less important for the moment, at least for you. Like, we can always just take the post-meeting notes and then add things onto linear there, as opposed to, like, hey, we need to do this live. And so I think what we talked about today is the sort of core things that we want to accomplish.
642 00:49:47.770 ⇒ 00:49:51.840 Uttam Kumaran: Okay. Okay, great. Yeah, and then if there’s any questions ever on, like.
643 00:49:52.220 ⇒ 00:49:54.819 Uttam Kumaran: hey, we use AI to create this update.
644 00:49:55.360 ⇒ 00:49:59.300 Uttam Kumaran: Do you think it’s too verbose or not? Like, those are simple things for me to answer, so…
645 00:49:59.430 ⇒ 00:50:04.700 Uttam Kumaran: I know that… that could be almost, like, Hard to isolate, so…
646 00:50:05.210 ⇒ 00:50:05.870 Samuel Roberts: Yes.
647 00:50:05.870 ⇒ 00:50:14.470 Uttam Kumaran: Cool. Okay, alright, let’s transition to the next stand up. And actually, I will have this open.
648 00:50:14.720 ⇒ 00:50:15.350 Samuel Roberts: Perfect.
649 00:50:15.350 ⇒ 00:50:19.420 Uttam Kumaran: On the side here. And then for context for everybody, this is…
650 00:50:19.690 ⇒ 00:50:23.209 Uttam Kumaran: what the AI team is working on in our new, sort of, like.
651 00:50:23.680 ⇒ 00:50:29.690 Uttam Kumaran: sprint-based delivery fashion. Team is working right now on how do we
652 00:50:30.280 ⇒ 00:50:37.980 Uttam Kumaran: Ship a stand-up assistant for us to run more efficient, more, accurate stand-ups.
653 00:50:38.300 ⇒ 00:50:41.859 Uttam Kumaran: As you can tell, like, we’re kinda gonna get to about…
654 00:50:42.120 ⇒ 00:50:46.969 Uttam Kumaran: 6 or 7 clients that need to get discussed in any stand-up, the best and the worst.
655 00:50:47.080 ⇒ 00:50:49.139 Uttam Kumaran: And it’s extremely hard.
656 00:50:49.760 ⇒ 00:50:54.680 Uttam Kumaran: maybe I don’t… I’m not sharing how hard it is. Extremely hard to…
657 00:50:54.820 ⇒ 00:50:58.589 Uttam Kumaran: go through every client and think about everything. And so, this is, like.
658 00:50:58.720 ⇒ 00:51:03.660 Uttam Kumaran: already gonna be very helpful. So, yeah, I’ll go ahead and share, and we can jump into the,
659 00:51:04.160 ⇒ 00:51:05.580 Uttam Kumaran: the next things.
660 00:51:07.510 ⇒ 00:51:11.559 Uttam Kumaran: Cool, so let’s talk about, insomnia first.
661 00:51:12.100 ⇒ 00:51:15.019 Uttam Kumaran: Maybe, Robert, you sent an update yesterday…
662 00:51:15.260 ⇒ 00:51:16.689 Uttam Kumaran: Do we want to start there?
663 00:51:18.180 ⇒ 00:51:23.790 Robert Tseng: Yeah, I…
664 00:51:29.450 ⇒ 00:51:33.000 Robert Tseng: I’m, like, trying to recall where we even were.
665 00:51:33.000 ⇒ 00:51:34.569 Uttam Kumaran: From… I’ll pull… I’ll pull it up.
666 00:51:46.200 ⇒ 00:51:49.420 Robert Tseng: Yeah, I mean, I think really what we need is just to get time
667 00:51:49.630 ⇒ 00:52:01.249 Robert Tseng: I mean, so two things. I would like to introduce Amber to the team. I’m still a bit hesitant, because, like, I feel like it may run into, like, a similar situation where she’s gonna…
668 00:52:01.300 ⇒ 00:52:11.139 Robert Tseng: They just make it look like we don’t really understand what’s going on, so I’ve been waiting for that, like, ramp-up period to kind of happen. I think it’s almost there, but…
669 00:52:11.230 ⇒ 00:52:31.060 Robert Tseng: really, I just need someone spending time with Bertie, and, like, asking her how she’s taking, like, the analysis that we’re sending to her, and the changes that she’s making. So, I mean, I can spend, like, an hour with her, but I think I already have an hour booked with Enrita later this week, and so I don’t really want to book, like, more than an hour of my time with them.
670 00:52:32.650 ⇒ 00:52:39.950 Uttam Kumaran: Yeah, I agree. I mean, I think the feedback yesterday, Amber, was mainly… I think there just has to be one more pass
671 00:52:40.150 ⇒ 00:52:42.909 Uttam Kumaran: Like, one more super polished pass, basically.
672 00:52:42.910 ⇒ 00:52:44.059 Amber Lin: Yeah.
673 00:52:44.060 ⇒ 00:52:49.519 Uttam Kumaran: But one way of also doing this is, like, I would suggest if it’s a deck, ask design.
674 00:52:49.820 ⇒ 00:52:57.950 Uttam Kumaran: Like, if it’s late nights, send it to Ann. If it’s afternoon, send it to Hannah. Okay. Have them help you…
675 00:52:58.180 ⇒ 00:53:01.260 Uttam Kumaran: They’re the people with, like, the best
676 00:53:01.780 ⇒ 00:53:08.899 Uttam Kumaran: like, polish. They’re the only… they’re the best Polish people here, you know, so I would ask them for help.
677 00:53:09.340 ⇒ 00:53:09.720 Uttam Kumaran: Yeah.
678 00:53:09.720 ⇒ 00:53:11.719 Amber Lin: Great, because I struggle with that, and that takes.
679 00:53:11.720 ⇒ 00:53:30.529 Uttam Kumaran: Me too, me too, I’m not good at that, and so that’s… they’re the… they’re my entire, like, that’s exactly how I… what I do. So, if you’re struggling there, and I’ll… I’ll… and I can let them know that… that we’re doing that, and so it… they should… they should be okay doing that. They… that’s what we do across a lot of different decks, and so…
680 00:53:30.680 ⇒ 00:53:37.139 Uttam Kumaran: that’s probably the last piece. I think, Robert, you’re right. One thing that’s going to happen…
681 00:53:37.780 ⇒ 00:53:40.380 Uttam Kumaran: One thing that’s risky here is that…
682 00:53:41.100 ⇒ 00:53:45.860 Uttam Kumaran: Birdie, and these guys are, like… fairly senior marketers.
683 00:53:45.970 ⇒ 00:53:55.149 Uttam Kumaran: And so what we don’t want to happen is that, Amber, you, like, go on one-on-one with them, they, like, spew a bunch of stuff, they’re like, what do you think? And you’re like, I don’t know, I’ve never done this.
684 00:53:55.270 ⇒ 00:54:02.230 Uttam Kumaran: And so, that’s what we want to sort of, like, try and… Mitigate, so we’ll have to…
685 00:54:02.780 ⇒ 00:54:04.980 Uttam Kumaran: Think about how best we do this.
686 00:54:06.130 ⇒ 00:54:17.130 Robert Tseng: Yeah, that’s why I haven’t brought anyone else on the team into the calls with the CEO, but, like, I feel like we’re at a point now where I do… we do need somebody who’s, like, daily or, like, frequently
687 00:54:17.940 ⇒ 00:54:24.540 Robert Tseng: talking to Birdie, because, like, I… I don’t… I can’t answer all of Amber’s questions, like, I don’t really exactly know what Birdie’s doing, so…
688 00:54:26.150 ⇒ 00:54:35.640 Uttam Kumaran: What do you think is the best way to facilitate it? I mean, I can… I mean, I’m happy to just hop on with Birdie, and me and Amber do it for a while until…
689 00:54:36.130 ⇒ 00:54:37.750 Uttam Kumaran: And we kind of suss it out.
690 00:54:38.080 ⇒ 00:54:45.900 Robert Tseng: I mean, I think… ideally, I could just book a call, and I can just sit in on the call as well, and
691 00:54:46.030 ⇒ 00:54:50.040 Robert Tseng: I mean, tomorrow is kind of their last working day, they don’t really do Friday, so…
692 00:54:50.210 ⇒ 00:54:53.570 Robert Tseng: I’m kind of booked out today, but,
693 00:54:55.010 ⇒ 00:55:09.909 Robert Tseng: nudge her, book a meeting with her tomorrow, and yeah, I guess we just… we actually have to run through what’s going on. We just have to run through what we’ve sent her, and, like, what… try to get her reactions and what she’s actually, using from what we’re sending her.
694 00:55:10.360 ⇒ 00:55:10.960 Uttam Kumaran: Okay.
695 00:55:11.800 ⇒ 00:55:14.010 Uttam Kumaran: I don’t mind doing that, and then let’s see.
696 00:55:14.140 ⇒ 00:55:16.450 Uttam Kumaran: Yeah, we don’t… I mean, what option do we have?
697 00:55:16.850 ⇒ 00:55:17.410 Robert Tseng: Yeah.
698 00:55:17.410 ⇒ 00:55:22.099 Uttam Kumaran: I can ping Birdie and say, like, hey, can we hop on? We can discuss these further.
699 00:55:22.650 ⇒ 00:55:28.149 Robert Tseng: Yeah, I think that’d be a good message to send today, and then just leave it open, like, we need to jump on today, we can do it.
700 00:55:28.150 ⇒ 00:55:29.340 Uttam Kumaran: Okay, great.
701 00:55:29.340 ⇒ 00:55:36.240 Robert Tseng: Yeah, I already gave Amber direction on, kind of, the next set of analysis last night, so I don’t think that we need to talk about that here.
702 00:55:36.240 ⇒ 00:55:39.209 Uttam Kumaran: Yeah, and I think the biggest thing for everybody here is, like.
703 00:55:39.780 ⇒ 00:55:47.770 Uttam Kumaran: There’s a lot of stuff that not everybody knows, but what… what kind of is important here is just poise under…
704 00:55:47.910 ⇒ 00:55:49.510 Uttam Kumaran: That type of pressure.
705 00:55:49.770 ⇒ 00:56:00.270 Uttam Kumaran: So, like, if a client asks, like, hey, like, have you guys… for example, I’m in the thing with, like, Hedra, and they’re like, have you used SIFT? Have you used all these things? I don’t know, I don’t know what those are.
706 00:56:00.340 ⇒ 00:56:13.390 Uttam Kumaran: what I do is I quickly Google it, and then I pattern match it to, something we’ve done, or I say, hey, let me go look into it, I’ll follow up with you. So you just find, like, the escape paths a little bit, and if that’s something that we need to maybe do.
707 00:56:13.850 ⇒ 00:56:23.299 Uttam Kumaran: something on, like a training internally, we can. I think the biggest risk we want to avoid is just, like, even if you don’t have the experience doing it.
708 00:56:23.660 ⇒ 00:56:24.520 Uttam Kumaran: don’t…
709 00:56:25.520 ⇒ 00:56:43.369 Uttam Kumaran: like, that is actually not relevant in that moment to the client. The client cares about the outcome. So it’s actually not that we don’t have experience in the one tool or the way, but it’s clear that we can do any analysis that they want. The more of what we need to do is just find ways to deflect. So that’s something that we can work on and talk through. If you’re in a moment like that, just let us know.
710 00:56:43.370 ⇒ 00:56:50.230 Uttam Kumaran: But that’s something I think… I’ll send a message today to Insomnia and try to get me and Amber on the phone with them, and yeah, we’ll see what happens.
711 00:56:50.230 ⇒ 00:56:51.010 Robert Tseng: Okay.
712 00:56:53.340 ⇒ 00:56:54.120 Uttam Kumaran: Cool.
713 00:56:55.010 ⇒ 00:56:59.970 Uttam Kumaran: Great, and then I saw a lot of your DBTs and PR reviews, so I can take a look at that.
714 00:57:00.460 ⇒ 00:57:03.140 Uttam Kumaran: Anything else?
715 00:57:03.360 ⇒ 00:57:04.640 Uttam Kumaran: on insomnia?
716 00:57:09.390 ⇒ 00:57:13.359 Uttam Kumaran: I, I still… I still yet haven’t gotten, like, approvals on…
717 00:57:13.740 ⇒ 00:57:24.839 Uttam Kumaran: stuff for infra, so I’m gonna basically message Robert and say… Robert Cantor and say, hey, can I… I can go ahead and kick off a Snowflake 30-day free trial, get Braze in there, and see what he says.
718 00:57:24.840 ⇒ 00:57:25.180 Robert Tseng: Yeah.
719 00:57:25.180 ⇒ 00:57:29.839 Uttam Kumaran: So, if he’s… I’m just gonna kinda just push.
720 00:57:30.050 ⇒ 00:57:30.650 Robert Tseng: Yeah.
721 00:57:30.770 ⇒ 00:57:32.300 Robert Tseng: Oh, that’s all we can do.
722 00:57:32.300 ⇒ 00:57:32.850 Uttam Kumaran: Okay.
723 00:57:33.140 ⇒ 00:57:42.559 Amber Lin: Robert, last question on Birdie, on Birdie’s questions. Is it all in the slide deck, or is there anything in Slack channels that I didn’t see?
724 00:57:44.190 ⇒ 00:57:51.510 Robert Tseng: There’s some stuff in Slack, but generally, she should… she has all her feedback in…
725 00:57:51.780 ⇒ 00:57:52.190 Amber Lin: Okay.
726 00:57:52.190 ⇒ 00:57:56.619 Robert Tseng: In the slide deck. I think her main concern, which I brought up to you yesterday, was, like.
727 00:57:57.140 ⇒ 00:58:07.929 Robert Tseng: No, I think it’s just the nuances of, like, audience size. We can’t just say, this campaign is better than that campaign. If you’re… if the sample size is, like, 10,000 versus
728 00:58:08.350 ⇒ 00:58:10.099 Robert Tseng: 500,000, it’s not a fair comparison.
729 00:58:10.100 ⇒ 00:58:12.490 Amber Lin: Yes. I hear you. Okay.
730 00:58:15.100 ⇒ 00:58:15.730 Uttam Kumaran: Okay.
731 00:58:16.850 ⇒ 00:58:19.500 Uttam Kumaran: Let’s talk about, Eden.
732 00:58:22.480 ⇒ 00:58:39.009 Robert Tseng: Yeah, Zora and I are on a lot of meetings today, where we have our calls with Attribution and with Northbeam, and then we already met on, kind of what we’re evaluating them on and, kind of the decision. I mean, I think by the end of the day, or by tomorrow, we should have a
733 00:58:39.080 ⇒ 00:58:43.770 Robert Tseng: Have a decision on which, which one, or what, what, what way to go forward.
734 00:58:43.860 ⇒ 00:58:53.620 Robert Tseng: I didn’t send De Milade’s dataset yet. I will do that today. I just… yeah, so that will clear up the…
735 00:58:59.710 ⇒ 00:59:03.930 Robert Tseng: Yeah, I forgot what we’ve been called, data recon work? Yeah.
736 00:59:06.110 ⇒ 00:59:15.329 Robert Tseng: Yeah, other than that, I feel like we’re pretty quiet, so I’m, like, mapping out some stuff to talk about in our, roadmap planning later today.
737 00:59:15.650 ⇒ 00:59:17.730 Uttam Kumaran: Okay, yeah, that’s, like, what I…
738 00:59:18.110 ⇒ 00:59:20.419 Uttam Kumaran: want to really bang out today, so…
739 00:59:20.420 ⇒ 00:59:20.840 Robert Tseng: Yeah.
740 00:59:20.840 ⇒ 00:59:27.470 Uttam Kumaran: For the folks joining that, right now I just have Henry Awash Robert,
741 00:59:27.950 ⇒ 00:59:29.970 Uttam Kumaran: maybe I can include,
742 00:59:30.320 ⇒ 00:59:37.419 Uttam Kumaran: Zoran, it’s kind of late your time, I mean, in case you’re up, you want to hang out, but maybe I’ll include Demulade, you there too, and…
743 00:59:37.770 ⇒ 00:59:43.319 Uttam Kumaran: You can just listen in, or… mainly, I want to go through Eden, Honey Stinger.
744 00:59:43.590 ⇒ 00:59:46.979 Uttam Kumaran: And insomnia as deep as we can go.
745 00:59:48.400 ⇒ 00:59:54.489 Uttam Kumaran: And maybe the goal is, like, I don’t know, guys, like, instead of planning… instead of grooming every client.
746 00:59:55.240 ⇒ 00:59:56.430 Uttam Kumaran: every week.
747 00:59:56.760 ⇒ 01:00:01.060 Uttam Kumaran: maybe we do this, we, like, batch 3, try to get 2 weeks ahead.
748 01:00:01.590 ⇒ 01:00:03.860 Uttam Kumaran: Next 3 next week, 2 weeks ahead.
749 01:00:05.700 ⇒ 01:00:10.009 Uttam Kumaran: I’m maybe just kind of talking to myself, thinking out loud about how to do this, but…
750 01:00:10.010 ⇒ 01:00:13.479 Amber Lin: It will work, because we don’t have time for more than 3 anyways.
751 01:00:13.480 ⇒ 01:00:14.920 Uttam Kumaran: Yeah, okay, okay.
752 01:00:15.030 ⇒ 01:00:19.670 Uttam Kumaran: Cool. So yeah, I think that’s the priority.
753 01:00:20.160 ⇒ 01:00:23.709 Uttam Kumaran: What else this week for, Eden?
754 01:00:24.070 ⇒ 01:00:26.970 Uttam Kumaran: Henry or… or Zoran.
755 01:00:31.530 ⇒ 01:00:37.619 Zoran Selinger: Attribution is really the main thing, and catalyst, we saw that, so,
756 01:00:37.830 ⇒ 01:00:49.370 Zoran Selinger: Yesterday, we… we did the implementation for Catalyst, so that’s… that’s ready. I think those two items are the… the biggest, from the… from my side for this week.
757 01:00:49.920 ⇒ 01:00:50.510 Uttam Kumaran: Okay.
758 01:00:52.460 ⇒ 01:01:04.079 Zoran Selinger: And they’re really happy with the solution, with saving the thank you page visits and attributing via that. They’re really, really happy about that. I think they’re going to… to…
759 01:01:04.080 ⇒ 01:01:06.989 Uttam Kumaran: Screenshots of them saying, I’m really happy about this.
760 01:01:06.990 ⇒ 01:01:11.260 Zoran Selinger: It was in the call with Ryan. It was in a call with Ryan, but…
761 01:01:11.260 ⇒ 01:01:15.129 Uttam Kumaran: The only thing is, like, those are very helpful for us to, like, save somewhere.
762 01:01:15.780 ⇒ 01:01:20.630 Uttam Kumaran: And I’m kind of, like, wanna keep those somewhere, so, okay, that’s fine, whatever.
763 01:01:20.850 ⇒ 01:01:23.210 Zoran Selinger: It was… Gotcha.
764 01:01:23.210 ⇒ 01:01:32.290 Uttam Kumaran: That’s for Eden, by the way, I’m not just picking on you. For any client, if they’re like, hey, we really like what you guys did, those were, like, amazing screenshots to just put in a folder, so, yeah.
765 01:01:33.640 ⇒ 01:01:34.300 Uttam Kumaran: Cool.
766 01:01:34.300 ⇒ 01:01:42.590 Zoran Selinger: Yeah, unfortunately, it was… it was a quick call. We try and… and he, he used… he used our word genius.
767 01:01:42.710 ⇒ 01:01:45.219 Zoran Selinger: Unfortunately, I don’t have that recording.
768 01:01:45.220 ⇒ 01:01:46.310 Uttam Kumaran: That’s okay, that’s okay.
769 01:01:46.310 ⇒ 01:01:50.500 Zoran Selinger: He’s really happy about it. I think they’re going to push, they’re really gonna push.
770 01:01:50.760 ⇒ 01:01:58.089 Zoran Selinger: Ingesting data with, with the attribution tool.
771 01:01:58.230 ⇒ 01:02:02.370 Zoran Selinger: Whatever we use, or… You know, a custom.
772 01:02:02.640 ⇒ 01:02:04.429 Zoran Selinger: So we’ll see, we’ll see.
773 01:02:05.300 ⇒ 01:02:05.900 Uttam Kumaran: Okay.
774 01:02:10.960 ⇒ 01:02:14.909 Amber Lin: Yesterday there was ad hoc dashboarding requests that came in.
775 01:02:15.090 ⇒ 01:02:21.680 Amber Lin: How much capacity do we have, and will Henry be able to take on that this week, or is this something we’re gonna push?
776 01:02:24.770 ⇒ 01:02:31.480 Uttam Kumaran: I guess it’s up to Henry. I know, Henry, you’re blocked on this, and you need some help, right? Last… yesterday, you mentioned you need some help from Awash.
777 01:02:31.940 ⇒ 01:02:32.440 Uttam Kumaran: I’d rather.
778 01:02:33.670 ⇒ 01:02:33.990 Uttam Kumaran: Yeah.
779 01:02:33.990 ⇒ 01:02:36.030 Henry Zhao: Yeah, I don’t think we’re blocked anymore, so we’re good.
780 01:02:36.440 ⇒ 01:02:37.000 Uttam Kumaran: Okay.
781 01:02:38.760 ⇒ 01:02:41.559 Henry Zhao: And yeah, Amber, I saw that task, already on my…
782 01:02:43.150 ⇒ 01:02:52.830 Uttam Kumaran: Yeah, I feel like we should move it in, that’s fine. I think, yeah, I feel like we should move it in. Right now, Henry’s mainly just focused on Eden, so I think this week we’re gonna plan out a few other…
783 01:02:53.470 ⇒ 01:02:57.410 Uttam Kumaran: a few other clients. I feel like what’s probably most relevant
784 01:02:57.670 ⇒ 01:03:01.199 Uttam Kumaran: Amber, in that… for that question, is, like, for us to basically be, like.
785 01:03:01.380 ⇒ 01:03:03.110 Uttam Kumaran: How are we doing on ours?
786 01:03:03.110 ⇒ 01:03:03.790 Amber Lin: Yeah.
787 01:03:03.830 ⇒ 01:03:06.149 Uttam Kumaran: And how are we doing on points?
788 01:03:06.270 ⇒ 01:03:08.730 Uttam Kumaran: So that’s something that I know we’re kind of, like.
789 01:03:08.970 ⇒ 01:03:14.350 Uttam Kumaran: there’s just so much stuff that came up for Eden that we didn’t end up, like, doing a lot of grooming.
790 01:03:14.680 ⇒ 01:03:16.680 Uttam Kumaran: But overall, like, I feel like…
791 01:03:17.330 ⇒ 01:03:22.520 Uttam Kumaran: If you can get that out, Henry, and it’s not, like, more than 2-3 hours, like, I feel okay about it.
792 01:03:23.680 ⇒ 01:03:24.660 Henry Zhao: Yeah.
793 01:03:26.160 ⇒ 01:03:27.270 Uttam Kumaran: Which one is that?
794 01:03:27.780 ⇒ 01:03:32.870 Amber Lin: I sent it in the channel as, 1114.
795 01:03:33.700 ⇒ 01:03:40.159 Amber Lin: Should be in to-do… Yeah, there we go.
796 01:03:40.260 ⇒ 01:03:41.360 Amber Lin: From Sarah.
797 01:03:43.300 ⇒ 01:03:44.010 Uttam Kumaran: Nice.
798 01:03:48.520 ⇒ 01:03:49.820 Uttam Kumaran: And who’s Sarah?
799 01:03:50.310 ⇒ 01:03:54.459 Amber Lin: Someone on their farm ops team, which means that
800 01:03:54.680 ⇒ 01:04:01.110 Amber Lin: It’s not the… it’s probably a request that trickle down from the top, and I feel like we’ve done very similar things before.
801 01:04:01.110 ⇒ 01:04:08.030 Uttam Kumaran: So, Robert, for these, like, do we have SLAs? Like, I would prefer to just be like, okay, we’re gonna tackle these next week.
802 01:04:08.460 ⇒ 01:04:10.950 Uttam Kumaran: by default, Care?
803 01:04:11.270 ⇒ 01:04:22.379 Robert Tseng: Yeah, well, I mean, her question is, like, do we have something that already exists? I think if there is, we should just give it to her immediately, or give a point or a direction, but anything that new, we don’t have to do immediately, yeah.
804 01:04:22.380 ⇒ 01:04:27.689 Uttam Kumaran: Okay. So, I guess, question back to Henry. Does… does this exist somewhere?
805 01:04:28.390 ⇒ 01:04:32.040 Henry Zhao: I would say no, but Robert, do you know of anything that exists already like this?
806 01:04:32.190 ⇒ 01:04:34.549 Robert Tseng: I’m not looking these days, I have no idea.
807 01:04:34.550 ⇒ 01:04:38.349 Uttam Kumaran: Yeah, it’s gonna be all between you, Awash, MLRA.
808 01:04:38.790 ⇒ 01:04:39.930 Henry Zhao: Yeah, but I’ll look into it.
809 01:04:40.290 ⇒ 01:04:43.890 Uttam Kumaran: So I would say, like, don’t wor- I guess my point is, like.
810 01:04:44.160 ⇒ 01:04:52.400 Uttam Kumaran: don’t work… like, if you’re gonna end up building something, pinging the channel, most likely I would like us to push this to next week.
811 01:04:52.540 ⇒ 01:04:53.239 Uttam Kumaran: This doesn’t seem.
812 01:04:53.240 ⇒ 01:04:53.580 Henry Zhao: Okay.
813 01:04:53.580 ⇒ 01:04:54.139 Uttam Kumaran: at all.
814 01:04:54.530 ⇒ 01:04:56.260 Henry Zhao: Sounds good. Okay. I agree.
815 01:04:57.480 ⇒ 01:04:58.820 Henry Zhao: Cool.
816 01:04:58.930 ⇒ 01:05:03.070 Uttam Kumaran: Okay, let’s go back…
817 01:05:03.910 ⇒ 01:05:08.000 Uttam Kumaran: So we feel good about Eden. Ellie, I…
818 01:05:08.270 ⇒ 01:05:15.720 Uttam Kumaran: I reviewed a little bit of your contract, I know you sent it, or the SOW. I can do another look today.
819 01:05:15.720 ⇒ 01:05:22.389 Robert Tseng: Yeah, all good, I already sent it to her. She poked around the doc. I haven’t followed up if she had questions, but yeah.
820 01:05:22.390 ⇒ 01:05:22.990 Uttam Kumaran: Okay.
821 01:05:23.220 ⇒ 01:05:24.530 Uttam Kumaran: Okay, great.
822 01:05:24.890 ⇒ 01:05:33.470 Uttam Kumaran: What’s, what’s the… plan for, README, I guess we’re meeting tomorrow.
823 01:05:34.610 ⇒ 01:05:38.440 Robert Tseng: I think the meeting got canceled tomorrow, for tomorrow. Okay.
824 01:05:39.270 ⇒ 01:05:44.959 Robert Tseng: Yeah, Icona was just like, I’m not gonna do anything this week. They didn’t give us any feedback, they’re on an off-site, so…
825 01:05:44.960 ⇒ 01:05:45.790 Uttam Kumaran: Okay, okay.
826 01:05:45.790 ⇒ 01:05:46.350 Robert Tseng: Yeah.
827 01:05:46.710 ⇒ 01:05:53.070 Uttam Kumaran: Alright, then I’m basically gonna move these all to… To do…
828 01:05:54.280 ⇒ 01:05:58.160 Uttam Kumaran: Okay, that’s helpful. Let’s see if… would this have caught that?
829 01:05:58.910 ⇒ 01:06:05.569 Uttam Kumaran: Okay, rame, Sonya, Hype, Eden, who are we missing?
830 01:06:09.150 ⇒ 01:06:15.649 Uttam Kumaran: I feel like that’s probably… it, right? We’ll do Urban STEM stand-up in the next meeting.
831 01:06:17.610 ⇒ 01:06:18.389 Amber Lin: We talked about…
832 01:06:18.390 ⇒ 01:06:22.600 Uttam Kumaran: of… Yeah, defaults, I feel like you both are…
833 01:06:22.600 ⇒ 01:06:24.690 Amber Lin: Do there, or is it just waiting for.
834 01:06:24.690 ⇒ 01:06:29.620 Uttam Kumaran: No, I would say just, like, if she doesn’t respond today, we’ll ping again.
835 01:06:29.620 ⇒ 01:06:29.990 Amber Lin: Okay.
836 01:06:29.990 ⇒ 01:06:33.059 Uttam Kumaran: But, kind of like, I think I have two…
837 01:06:33.180 ⇒ 01:06:37.910 Uttam Kumaran: questions for everybody. One is, like, we’re starting to ship
838 01:06:38.590 ⇒ 01:06:42.320 Uttam Kumaran: Analysis pretty consistently across a few clients.
839 01:06:42.520 ⇒ 01:06:49.219 Uttam Kumaran: I would really like us to get in the habit of, like, getting reviews from, like, Everybody, if possible.
840 01:06:49.380 ⇒ 01:06:57.030 Uttam Kumaran: Would… would it be fair to be like, hey, if you do a… if you do an analysis, share it in, like, Data Team?
841 01:06:57.140 ⇒ 01:07:06.890 Uttam Kumaran: And then be like, hey, can I get comments? Like, right now, I think comments are really constrained to probably just people who are on the client team, if that. It’s probably mostly Robert and I.
842 01:07:07.600 ⇒ 01:07:14.780 Uttam Kumaran: I would like, Amber, like, for example, if you sent that pricing thing, you could just send it to the data team, be like, here, I would just love our lies on this.
843 01:07:15.280 ⇒ 01:07:16.540 Amber Lin: Yeah.
844 01:07:16.740 ⇒ 01:07:21.990 Amber Lin: Totally, I don’t mind, it’s just as… I hope people look at it. I will love some more feedback.
845 01:07:22.270 ⇒ 01:07:23.620 Uttam Kumaran: Okay, great.
846 01:07:23.740 ⇒ 01:07:32.289 Uttam Kumaran: Cool. And then, yeah, we have, this client, Hedra, that started today. I will… I can…
847 01:07:32.560 ⇒ 01:07:39.509 Uttam Kumaran: I’ll share a little bit more about it, but probably on Monday, and then, yeah, Honey Singer also, we’ll plan later today.
848 01:07:39.710 ⇒ 01:07:44.120 Uttam Kumaran: Anything else for us to… Discuss?
849 01:07:50.070 ⇒ 01:07:50.710 Robert Tseng: Nope.
850 01:07:53.760 ⇒ 01:07:58.409 Uttam Kumaran: Great, and then, okay, I feel like that’s… That’s kind of it.
851 01:08:00.010 ⇒ 01:08:03.499 Uttam Kumaran: I don’t think what I wanted to talk about, if anything’s missing.
852 01:08:06.400 ⇒ 01:08:13.369 Uttam Kumaran: Yeah, I think the only other thing for Urban Stems, Robert, is I need some help on analysis. Maybe next week I can grab time with you.
853 01:08:13.570 ⇒ 01:08:14.270 Robert Tseng: Yeah.
854 01:08:14.270 ⇒ 01:08:16.039 Uttam Kumaran: What do you think is best, like…
855 01:08:16.390 ⇒ 01:08:21.080 Uttam Kumaran: Do you wanna do, like, an hour of, like, Discovery, and then…
856 01:08:21.510 ⇒ 01:08:24.120 Uttam Kumaran: We can pass something to the team to help.
857 01:08:24.540 ⇒ 01:08:26.130 Uttam Kumaran: Do an analysis on.
858 01:08:27.540 ⇒ 01:08:33.250 Uttam Kumaran: It’s, again, both on forecasting side and on… Shipping.
859 01:08:36.060 ⇒ 01:08:51.690 Robert Tseng: Yeah, I guess I haven’t looked at Urban STEM’s data in a long time, so… Yeah. Kind of just, yeah, whatever context you have on, like, what they’ve already shared. I mean, I basically… yeah, I’ll just do a discovery by poking around in there, and kind of like how I build out the roadmaps for… for the other clients.
860 01:08:52.189 ⇒ 01:08:53.329 Uttam Kumaran: Okay, okay, cool.
861 01:08:54.189 ⇒ 01:08:55.479 Uttam Kumaran: Okay, great.
862 01:08:55.849 ⇒ 01:09:04.539 Uttam Kumaran: Alright guys, well, message me if anything, maybe Urban STEM’s crew, we can stay on here for another 5 minutes, and I can just talk briefly before the next stand-up.
863 01:09:06.579 ⇒ 01:09:07.739 Uttam Kumaran: Thank you, everyone.
864 01:09:07.740 ⇒ 01:09:08.420 Robert Tseng: Yep.
865 01:09:08.420 ⇒ 01:09:09.129 Zoran Selinger: Thanks, Matt.
866 01:09:09.130 ⇒ 01:09:09.520 Uttam Kumaran: Bye.
867 01:09:09.529 ⇒ 01:09:10.139 Samuel Roberts: Wow.
868 01:09:10.729 ⇒ 01:09:11.670 Henry Zhao: Thanks, guys.
869 01:09:11.950 ⇒ 01:09:12.660 Uttam Kumaran: Thanks.
870 01:09:14.870 ⇒ 01:09:25.500 Uttam Kumaran: Cool, so in terms of this week, I’m gonna have a little bit of time later today to do some Looker work. Kind of just two things I wanted to flag.
871 01:09:25.840 ⇒ 01:09:30.840 Uttam Kumaran: I sent this PR last week,
872 01:09:31.500 ⇒ 01:09:36.960 Uttam Kumaran: It’s kind of like a big, like, optimization PR?
873 01:09:37.950 ⇒ 01:09:39.040 Uttam Kumaran: I guess, like…
874 01:09:40.080 ⇒ 01:09:45.959 Uttam Kumaran: I mean, we’re still in the point where I think Emily’s just spazzing PRs into there, so I’m kind of like…
875 01:09:46.120 ⇒ 01:09:54.919 Uttam Kumaran: I don’t personally want to break things, but I did a lot of performance optimization on some of these tables, and added a lot of helpful context.
876 01:09:55.140 ⇒ 01:09:56.930 Uttam Kumaran: Can I just, like…
877 01:09:58.290 ⇒ 01:10:05.250 Uttam Kumaran: get eyes on this, or I guess you guys both approved it. I’ll just review this, but if you guys are okay, I’ll push this in today.
878 01:10:05.350 ⇒ 01:10:10.960 Uttam Kumaran: just… I just want to make sure everybody’s, like, online in case this, like, some… one of these tables, like, fucks up.
879 01:10:13.050 ⇒ 01:10:13.780 Awaish Kumar: Okay, yeah.
880 01:10:13.780 ⇒ 01:10:14.560 Demilade Agboola: Yeah.
881 01:10:16.870 ⇒ 01:10:19.270 Uttam Kumaran: Okay, cool. And then,
882 01:10:19.780 ⇒ 01:10:26.449 Uttam Kumaran: Yeah, I know… I think, Awash, like, this is probably good for us… time for us to really stress test Metaplane.
883 01:10:26.640 ⇒ 01:10:30.500 Uttam Kumaran: Like, if it’s not doing what we need to do, then we can consider another tool.
884 01:10:30.870 ⇒ 01:10:33.159 Uttam Kumaran: I feel like it’s okay.
885 01:10:33.450 ⇒ 01:10:36.709 Uttam Kumaran: But, like, I kind of don’t like that they do this, like, detection thing.
886 01:10:36.890 ⇒ 01:10:39.739 Uttam Kumaran: Like, why can’t we just put SLAs on tables, you know?
887 01:10:42.140 ⇒ 01:10:54.769 Awaish Kumar: Yeah, like, we can basically, but we need to know then, like… so what right now we are doing is that it just trains itself on the data and figures out what is normal.
888 01:10:54.770 ⇒ 01:10:55.230 Uttam Kumaran: Okay.
889 01:10:55.550 ⇒ 01:11:02.509 Awaish Kumar: But we can, basically, if we know by each table, 2 over… 2 hour, like, is minimum SLA for…
890 01:11:02.720 ⇒ 01:11:06.289 Awaish Kumar: For this field, or for this table, yeah, we can add that.
891 01:11:06.620 ⇒ 01:11:12.270 Uttam Kumaran: Okay, so next, so, when we get into the next meeting, let’s talk about that really quick, and get a decision from Emily.
892 01:11:13.200 ⇒ 01:11:14.749 Uttam Kumaran: And let’s put SLAs.
893 01:11:15.590 ⇒ 01:11:19.709 Uttam Kumaran: I would like us to do at least 4… I mean, what do you think, Devin, like…
894 01:11:19.840 ⇒ 01:11:27.160 Uttam Kumaran: I would say at least… 4 hours for all core MARTS models, right? At least Monday to Friday. Yeah.
895 01:11:29.660 ⇒ 01:11:31.500 Demilade Agboola: Yeah.
896 01:11:33.200 ⇒ 01:11:34.030 Uttam Kumaran: And then the XFJ.
897 01:11:34.030 ⇒ 01:11:37.709 Demilade Agboola: We definitely have… yeah, the XF tables, we should try and get that.
898 01:11:38.150 ⇒ 01:11:40.510 Demilade Agboola: I think we’re just in that phase where…
899 01:11:41.950 ⇒ 01:11:47.520 Demilade Agboola: We don’t want to invest so much into, like, maintaining those tables, because we’re literally.
900 01:11:47.520 ⇒ 01:11:52.969 Uttam Kumaran: No, totally, but I just want to make sure that we get flagged first.
901 01:11:53.250 ⇒ 01:11:55.780 Uttam Kumaran: If they’re not running. Versus…
902 01:11:55.780 ⇒ 01:11:56.350 Demilade Agboola: Yeah.
903 01:11:56.350 ⇒ 01:11:57.699 Uttam Kumaran: Like this, you know?
904 01:11:58.430 ⇒ 01:12:06.720 Demilade Agboola: Yeah, definitely. We could definitely put, like, a 3-hour… a 3-hour SLA, based off how often we expect it to run.
905 01:12:06.880 ⇒ 01:12:08.879 Uttam Kumaran: Okay. Especially during business hours.
906 01:12:09.390 ⇒ 01:12:10.370 Uttam Kumaran: Okay, okay.
907 01:12:11.120 ⇒ 01:12:18.389 Uttam Kumaran: And my other question, and this is probably… yeah, my other question, and this is maybe a waste since you have this ticket.
908 01:12:18.550 ⇒ 01:12:24.120 Uttam Kumaran: The sta- I still think that, like, I don’t know if you had put any thought into, like, how we’re gonna do staging.
909 01:12:24.690 ⇒ 01:12:27.120 Uttam Kumaran: Did you end up thinking about that at all?
910 01:12:28.850 ⇒ 01:12:31.680 Awaish Kumar: First, yeah, like, I added it…
911 01:12:32.160 ⇒ 01:12:35.320 Awaish Kumar: figure just today for, like, DBT jobs.
912 01:12:35.600 ⇒ 01:12:36.650 Awaish Kumar: So…
913 01:12:37.260 ⇒ 01:12:47.629 Uttam Kumaran: I… yeah, to give you my perspective, I feel like it’s not going to be possible to create a single schema for all staging jobs in dbt Cloud.
914 01:12:48.310 ⇒ 01:12:49.900 Uttam Kumaran: I did some digging.
915 01:12:50.010 ⇒ 01:12:51.969 Uttam Kumaran: Like, a couple months ago?
916 01:12:52.650 ⇒ 01:12:58.810 Uttam Kumaran: my alternative… solution is… do it in GitHub Actions.
917 01:12:59.110 ⇒ 01:13:03.610 Uttam Kumaran: run a staging job on all PRs, remove the dbt cloud staging.
918 01:13:04.200 ⇒ 01:13:09.520 Uttam Kumaran: That way, I actually want to create staging, like, Looker Explores.
919 01:13:09.940 ⇒ 01:13:12.929 Uttam Kumaran: So that we can go check the data.
920 01:13:13.050 ⇒ 01:13:20.840 Uttam Kumaran: Right? Like, before, like, we can actually do end-to-end testing.
921 01:13:21.600 ⇒ 01:13:26.559 Uttam Kumaran: on staging. Right now, it’s, like, really, really not possible. Additionally.
922 01:13:26.670 ⇒ 01:13:36.009 Uttam Kumaran: We’re having the issue where every… because all these schemas are getting created, the incremental jobs all have to run a full refresh the moment they get affected.
923 01:13:36.150 ⇒ 01:13:38.880 Uttam Kumaran: And so, like, the PRs never run, they all time out.
924 01:13:40.090 ⇒ 01:13:46.170 Uttam Kumaran: So that staging schema we can just run at least once a day, if not on every PR.
925 01:13:46.860 ⇒ 01:13:48.559 Uttam Kumaran: Is my thought.
926 01:13:48.920 ⇒ 01:13:50.689 Uttam Kumaran: That’s how I’ve done it in the past.
927 01:13:55.500 ⇒ 01:14:02.059 Awaish Kumar: Okay, yeah, like… Give me today, and I can…
928 01:14:02.550 ⇒ 01:14:03.130 Uttam Kumaran: Okay.
929 01:14:04.480 ⇒ 01:14:08.069 Uttam Kumaran: That would be very, very nice. Okay, cool. Alright, let’s,
930 01:14:09.140 ⇒ 01:14:10.490 Uttam Kumaran: We’ll hop to the next one.
931 01:14:12.370 ⇒ 01:14:13.169 Uttam Kumaran: Thank you, guys.